Apostol Solutions Vol2
Apostol Solutions Vol2
SOLUTIONS TO CALCULUS VOLUME 2 MULTI-VARIABLE CALCULUS AND LINEAR ALGEBRA, WITH
APPLICATIONS TO DIFFERENTIAL EQUATIONS AND PROBABILITY BY TOM APOSTOL
ERNEST YEUNG - PRAHA 10, ČESKÀ REPUBLIKA
1.5 E XERCISES - I NTRODUCTION , T HE DEFINITION OF A LINEAR SPACE , E XAMPLES OF LINEAR SPACES ,
E LEMENTARY CONSEQUENCES OF THE AXIOMS
Recall the following:
Definition 1 (Linear Space.).
Let V be a nonempty set of objects.
Linear space if a set V that satisfies the following ten axioms.
(1) (closure under addition) ∀x, y ∈ V, x + y ∈ V
(2) (closure under scalar multiplication) ∀x ∈ V, αx ∈ V
(3) (Additive commutativity) ∀x, y ∈ V, x + y = y + x
(4) (Additive Associativity) ∀x, y ∈ V, (x + y) + z = x + (y + z)
(5) (Additive Identity Existence) ∃0 ∈ V such that
x + 0 = x, ∀x ∈ V
(6) (Additive Inverse Existence) ∃(−1)x such that
x + (−1)x = 0
(7) (Scalar Associativity) ∀x ∈ V, ∀α, β ∈ R or α, β ∈ C
(αβ)x = α(βx)
(8) (distributivity for addition in V ) ∀x, y ∈ V ; ∀a ∈ R or ∀a ∈ C,
a(x + y) = ax + ay
(9) (distributivity for addition of numbers) ∀x ∈ V, ∀a, b ∈ R or ∀a, b ∈ C,
(a + b)x = ax + bx
(10) (Multiplicative identity existence) ∀x ∈ V, 1x = x
p r
Exercise 1. Consider x = q,y = s ∈ V where p, q, r, s are polynomials. ps + rq, qs are polynomials as well.
p r ps + rq
x+y =+ = ∈V
q s qs
α ∈ R, αp is a polynomial
αp
αx = ∈V
q
ps + rq rq + ps
x+y = = =r+p
qs qs
p r t p r t
(x + y) + z = + + = + + = x + (y + z)
q s v q s v
p 0 p 0
x + 0 = + = = x so ∈ V if q 6= 0
q q q q
p p 0
x + (−1)x = + (−1) = = 0
q q q
(αβ)p α(βp)
(αβ)x = = = α(βx) (follows from associativity of real or complex numbers)
q q
α(x + y) = αx + αy and (α + β)x = αx + βx follows from distributivity for real numbers
p q p q
Consider x = = = (1)x, ∈ V
q q q q
1
Exercise 3. All f with f (0) = f (1)
f (0) + g(0) = (f + g)(0) = f (1) + g(1) = (f + g)(1)
af (0) = (af )(0) = af (1) = (af )(1)
f (x) + g(x) = (f + g)(x) = g(x) + f (x) = (g + f )(x)
(f (x) + g(x)) + h(x) = ((f + g) + h)(x) = f (x) + (g(x) + h(x)) = (f + (g + h))(x)
0(x) = 0 (f + 0)(x) = f (x) + 0(x) = f (x)
(−1)f (x) = (−f )(x) (f + (−1)f )(x) = (f − f )(x) = f (x) + (−1)f (x) = 0 = 0(x)
(αβ)f (x) = α(βf (x)) = α(βf )(x)
a(f + g)(x) = a(f (x) + g(x)) = af (x) + ag(x) = (af + ag)(x)
(a + b)f (x) = af (x) + bf (x) = (af )(x) + (bf )(x) = (af + bf )(x)
(1f )(x) = f (x)
Exercise 4. All f with 2f (0) = f (1)
2(f + g)(0) = 2f (0) + 2g(0) = f (1) + g(1) = (f + g)(1)
2(αf )(0) = 2αf (0) = αf (1) = (αf )(1)
f + g = g + f, (f + g) + h = f + (g + h) follow from properties of the reals.
20(0) = 0 = 0(1); (f + 0)(x) = f (x) + 0 = f (x)
(f + (−f ))(x) = f (x) + −f (x) = 0
(αβ)f (x) = α(βf (x))
a(f + g)(x) = (af )(x) + (ag)(x), (a + b)f (x) = af (x) + bf (x) follow from properties of the reals.
1x = x
Exercise 5. All f with f (1) = 1 + f (0)
f (1) + g(1) = (f + g)(1) = 1 + f (0) + 1 + g(0) = 2 + (f + g)(0)
So closure under addition is violated.
Exercise 6. All step functions defined on [0, 1].
Exercise 8. Even functions, f (−x) = f (x).
(f + g)(−x) = f (−x) + g(−x) = f (x) + g(x)
(αf )(−x) = αf (−x) = αf (x) = (αf )(x)
(f + g)(x) = f (x) + g(x) = g(x) + f (x) = (g + f )(x) (follows from commutativity of real numbers)
((f + g) + h)(−x) = (f + g)(−x) + h(−x) = f (x) + g(x) + h(x) = f (x) + (g + h)(x) = (f + (g + h))(x)
(follows from associativity of real numbers)
if 0(−x) = 0(x) = 0 ∀x ∈ D so 0 exists and (f + 0)(x) = f (x) + 0(x) = f (x)
(−f )(−x) = −f (−x) = −f (x) so (f + −f )(x) = f (x) − f (x) = 0(x)
α(βf )(−x) = α(β(f (−x))) = α(βf (x)) = α(βf )(x)
((αβ)f )(−x) = (αβ)f (−x) = (αβ)f (x) = ((αβ)f )(x) (from associativity of the real numbers)
α(f + g)(x) = α(f (x) + g(x)) = (αf )(x) + (αg)(x) = (αf + αg)(x)
(α + β)f (x) = αf (x) + βf (x) = (αf )(x) + (βf )(x) = (αf + βf )(x)
1f (x) = 1(f (x)) = f (x)
Exercise 22. All vectors (x, y, z) in V3 with z = 0.
This space is closed under addition and scalar multiplication since
(x1 , y1 , 0) + (x2 , y2 , 0) = (x1 + x2 , y1 + y2 , 0)
k(x, y, 0) = (kx, ky, 0)
both belong to this space.
Additive commutativity, additive associativity, scalar associativity, distributivity in addition in V , and distributivity in addition
of numbers are satisfied automatically, since all vectors in this subset belong to V3 , a linear space.
(0, 0, 0) belongs to this space, since z = 0, so existence of an additive identity is fulfilled.
2
(−x, −y, 0) = −(x, y, 0) belongs in this space and so the existence of an additive inverse for each element (x, y, 0) in this
space is fulfilled.
1(x, y, 0) = (x, y, 0), and so multiplicative identity existence is fulfilled.
This is a linear space. Note that we could’ve also said that this space is exactly V2 , and V2 is a linear space.
Exercise 23. All vectors (x, y, z) in V3 with x = 0 or y = 0.
Consider (0, y1 , z1 ) + (x2 , 0, z2 ) = (x2 , y1 , z1 + z2 ). This vector does not belong to this space. This is not a linear space.
Exercise 24. All vectors (x, y, z) in V3 with y = 5x.
This space is closed under addition and scalar multiplication since
(x1 , 5x1 , 0) + (x2 , 5x2 , 0) = (x1 + x2 , 5x1 + 5x2 , 0) =⇒ 5x1 + 5x2 = 5(x1 + x2 )
k(x1 , 5x1 , z1 ) =⇒ k5x1 = 5(kx1 )
Additive commutativity, additive associativity, scalar associativity, distributivity in addition in V , and distributivity in addi-
tion of numbers are satisfied automatically, since all vectors in this subset belong to V3 , a linear space.
(0, 5(0), 0) belongs to this space, so existence of an additive identity is fulfilled.
(−x, 5(−x), −z) = −(x, 5x, z) belongs in this space and so the existence of an additive inverse for each element in this
space is fulfilled.
1(x, 5x, z) = (x, 5x, z), and so multiplicative identity existence is fulfilled.
This is a linear space.
Exercise 25. All vectors (x, y, z) in V3 with 3x + 4y = 1 z = 0.
Consider closure:
(x1 , y1 , 0) + (x2 , y2 , 0) = (x1 + x2 , y1 + y2 , 0)
z3 = 0; 3(x1 + x2 ) + 4(y1 + y2 ) = 2
Closure under addition is not satisfied. Thus, this is not a linear space.
Exercise 26. All vectors (x, y, z) in V3 which are scalar multiples of (1, 2, 3).
Closure is fulfilled since x + y = a(1, 2, 3) + b(1, 2, 3) = (a + b)(1, 2, 3), which is a scalar multiple of (1, 2, 3), and
ax = ab(1, 2, 3), which is another scalar multiple of (1, 2, 3).
Additive commutativity, additive associativity, scalar associativity, distributivity in addition in V , and distributivity in addi-
tion of numbers are satisfied automatically, since all vectors in this subset belong to V3 , a linear space.
∃ 0 since 0(1, 2, 3) = 0 is a scalar multiple of (1, 2, 3).
∃ − 1x ∀ x, since (−1)a(1, 2, 3) is a scalar multiple of (1, 2, 3).
1(a(1, 2, 3)) = (a(1, 2, 3)) is a scalar multiple of (1, 2, 3) and so multiplicative identity existence is satisfied.
This is a linear space.
Exercise 28. All vectors in Vn that are linear combinations of 2 given vectors A and B.
a1 A + b1 B + a2 A + b2 B = (a1 + a2 )A + (b1 + b2 )B belongs in this space.
c(aA + bB) = (ca)A + (cb)B belongs in this space.
x + y = y + x, (x + y) + z = x + (y + z) follow since the vectors belong in Vn , a linear space.
0A + 0B = 0; a1 A + b1 B + 0A + 0B = a1 A + b1 B
−a1 A + −b1 B belongs in this space and a1 A + b1 B + −a1 A + −b1 B = (a1 − a1 )A + (b1 − b1 )B = 0
(ab)x = a(bx), a(x + y) = ax + ay, (a + b)x = ax + bx, 1x = x follow since the vectors in this space belong in Vn , a
linear space.
Exercise 29. Let V = R+ , let x“ +00 y = xy and a“ ·00 x = xa
x“ +00 y = xy ∈ R+
a“ ·00 x = xa = ea ln x > 0 a“ ·00 x ∈ R+
x“ +00 y = xy = yx = y“ +00 x
(x“ +00 y)“ +00 z = xyz = x(yz) = x“ +00 (y“ +00 z)
x“ +00 “000 = x1 = x so 00 000 = 1 ∈ R+
3
00
(−1)x00 = 1
x ∈ R+ x“ +00 “(−1)x00 = x x1 = 1 = “000
(ab)x = xab = (xb )a = a(bx)
a“ ·00 (x“ +00 y) = a“ ·00 (xy) = (xy)a = xa y a = (a“ · x)“ +00 (a“ ·00 y)
(a + b)“ ·00 x = xa+b = xa xb = xa “ +00 xb = a“ ·00 x“ +00 b“ ·00 x
1“ ·00 x = x1 = x
This is indeed a linear space.
Exercise 30.
(1) From Axiom 5,6, the Additive Identity Existence and Additive Inverse Existence, that ∃ 0 ∈ V s.t. x + 0 = x, ∀ x ∈
V and ∃ (−1)x s.t. x + (−1)x = 0, then, using associativity, commutativity, and distributivity for addition of
numbers,
x + 0 = x = x + (x + (−1)x) = 2x + (−1)x = (2 + (−1))x = 1x
(2) If Ax.6 is replaced by Ax.6’, ∀ x ∈ V, ∃ y ∈ V s.t. x + y = 0,
x + 0 = x = x + (x + y) = 2x + y = x
So Ax.10 does not hold since 2x + y = x.
Exercise 31.
(1) (x1 , x2 ) + (y1 , y2 ) = (x1 + y1 , x2 + y2 ) a(x1 , x2 ) = (ax1 , 0).
Not a linear space: violates Additive Inverse existence, which demands ∃ (−1)x s.t. x + (−1)x = 0, not ∃ y s.t.
x + y = 0, so that for (−1)x = (−x1 , 0), (x1 , x2 ) + (−1)x = (0, x2 ) 6= 0 and Multiplicative identity existence,
since 1x = (x1 , 0) 6= x.
(2) (x, 1, x2 ) + (y1 , y2 ) = (x1 + y1 , 0) a(x1 , x2 ) = (ax1 , ax2 )
Not a linear space: violates distributivity in addition of numbers, because
(a + b)x = ((a + b)x1 , (a + b)x2 ) but ax + bx = (ax1 + bx1 , 0)
(3) (x, 1, x2 ) + (y1 , y2 ) = (x1 , x2 + y2 ) a(x1 , x2 ) = (ax1 , ax2 )
Not a linear space because it violates Additive Commutativity of elements. (x1 , x2 ) + (y1 , y2 ) = (x1 , x2 + y2 ) but
(y1 , y2 ) + (x1 , x2 ) = (y1 , x2 + y2 )
(4) (x, 1, x2 ) + (y1 , y2 ) = (|x1 + y1 |, |y1 + y2 |) a(x1 , x2 ) = (|ax1 |, |ax2 |)
Not a linear linear space because it violates Distributivity for addition of numbers:
(a + b)(x1 , x2 ) = (|(a + b)x1 |, |(a + b)x2 |)
ax + bx = (|ax1 |, |ax2 |) + (|bx1 |, |bx2 |)
but |(a + b)x1 | ≤ |ax1 | + |bx1 | in general
Exercise 32. Theorem 1.3.
(1) 0x = 0
(2) a0 = 0
(3) (−a)x = −(ax) = a(−x)
(4) If ax = 0, then either a = 0 or x = 0
(5) If ax = ay and a 6= 0, then x = y
(6) If ax = bx and x 6= 0, then a = b
(7) −(x + y) = (−x) + (−y) = −x Pn− y
(8) x + x = 2x, x + x + x = 3x, j=1 x = nx
Part (d), or part 4, is proven by considering this:
If ax = 0,
then if a = 0 and x = 0, done.
If a 6= 0, ax + a0 = a(x + 0) = 0
since a is a real number, ∃ a1 ∈ R s.t. a1 a = 1
1
=⇒ 1(x + 0) = x + 0 = x = 0=0
a
If x 6= 0, suppose a 6= 0.
1 1
ax = 1x = x = 0 = 0
a a
But 0 is unique. Contradiction. So a = 0 and that’s okay, since by part (a) or part (1) of Thm. 1.3., 0x = 0.
4
For part (e), or part 5, if ax = ay, a 6= 0, then subtract ay from both sides to get ax − ay = 0 = a(x − y) = 0. Use
distributivity to get ax − ay = a(x − y) = 0. Since a 6= 0, then from part (d) or part 4, x − y = 0 must be true. Then
−y = −x or, multiplying both sides by −1, y = x.
For part (f), or part 6, if ax = ay, subtract bx from both sides and use distributivity to get ax − bx = (a − b)x = 0. Since
x 6= 0, then by part (d), or part 4, a − b = 0. Add b to both sides to get a = b.
For part (g), or part 7, note that from the existence of an additive inverse, x + −x = 0. Consider x + (−1)x = 0. x = 1x by
the existence of a multiplicative identity, and so using distributivity, 1x + (−1)x = (1 + −1)x = 0x = 0. Then (−1)x is also
an additive inverse for all x ∈ V . But additive inverses are unique, by theorem, so (−1)x = −x. Using that and distributivity,
we get (−x) + (−y) = (−1)x + (−1)y = (−1)(x + y) = −(x + y). −x − y = −(x + y) because
x + y + −(x + y) = 0 = x − x + y − y = x + y − x − y, where we used additive commutativity at the last step. Then
subtract x + y from both sides to get −x − y = −(x + y).
Pn identity and distributivity to get x + x = 1x + 1x = (1 + 1)x = 2x.
For part (h), or part 8, use the existence of a multiplicative
Now, we’ll use induction. Assume the nth case, that j=1 x = nx.
Pn+1 Pn+1 Pn
Consider j=1 x. j=1 x = j=1 x + x = nx + x = nx + 1x = (n + 1)x. Done.
1.10 E XERCISES - S UBSPACES OF A LINEAR SPACE , D EPENDENT AND INDEPENDENT SETS IN A LINEAR SPACE ,
BASES AND DIMENSION , C OMPONENTS
Exercise 1. x = 0
(0, y1 , z1 ) + (0, y2 , z2 ) = (0, y1 + y2 , z1 + z2 ) ∈ S
k(0, y, z) = (0, ky, kz) ∈ S
Yes, S is a subspace.
(0, y, z) = y(0, 1, 0) + z(0, 0, 1) ∈ S
0 = y(0, 1, 0) + z(0, 0, 1) =⇒ z = 0 y=0
dimS = 2
Exercise 2. x + y = 0
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since k(x, y, z) ∈ S since
x1 + x2 + y1 + y2 = 0 + 0 = 0 kx + ky = k(x + y) = 0
Yes, S is a subspace.
(x, y, z) = (x, −x, z) = x(1, −1, 0) + z(0, 0, 1)
0 = x(1, −1, 0) + z(0, 0, 1) =⇒ z = 0, x = 0
dimS = 2
Exercise 3. x + y + z = 0
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since k(x, y, z) ∈ S since
x1 + x2 + y1 + y2 + z1 + z2 = 0 + 0 = 0 k(x + y + z) = kx + ky + kz = 0
Yes, S is a subspace.
(x, y, −(x + y)) = x(1, 0, −1) + y(0, 1, −1)
0 = x(1, 0, −1) + y(0, 1, −1) =⇒ x = 0, y = 0
dimS = 2
Exercise 4. x = y
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since k(x, y, z) ∈ S since
x1 + x2 = y + 1 + y + 2 kx = ky
Yes, S is a subspace.
(x, y, z) = (x, x, z) = x(1, 1, 0) + z(0, 0, 1)
0 = x(1, 1, 0) + z(0, 0, 1) =⇒ x = 0, z = 0
dimS = 2
Exercise 5. x = y = z
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since k(x, y, z) ∈ S since
x1 + x2 = y + 1 + y + 2 = z1 + z2 kx = ky = kz
5
Yes, S is a subspace.
(x, y, z) = x(1, 1, 1)
0 = x(1, 1, 1) =⇒ x = 0,
dimS = 1
Exercise 6. x = y or x = z
If x1 + y1
If x2 = y2 , x1 + x2 = y1 + y2
x1 + x2 may not equal y1 + y2
else if x2 = z2 , No, S is not a subspace.
x1 + x2 may not equal z1 + z2
Exercise 7. x2 − y 2 = 0
(x1 + x2 )2 − (y12 + y22 ) = 2x1 x2 − 2y1 y2 maybe not equal to zero
S not a subspace.
Exercise 8. x + y = 1
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈
/ S since
x1 + x2 + y1 + y2 = 2
No S is not a subspace.
Exercise 9. y = 2x and z = 3x
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since (kx, ky, kz) ∈ S since
y1 + y2 = 2x1 + 2x2 = 2(x1 + x2 ) ky = k2x = 2kx
z1 + z2 = 3x1 + 3x2 = 3(x1 + x2 ) kz = k3x = 3kx
S is a subspace.
(x, y, z) = x(1, 2, 3)
0 = x(1, 2, 3) =⇒ x = 0
dimS = 1
Exercise 10.
x+y+z =0 x=0
⇐⇒
x−y−z =0 y = −z
(x1 + x2 , y1 + y2 , z1 + z2 ) ∈ S since
x1 + x2 = 0
y(0, 1, −1) = (x, y, z)
y1 + y2 = −z1 − z2 = −(z1 + z2 )
0 = y(0, 1, −1) =⇒ y = 0
k(x, y, z) ∈ S since
kx = 0 ky = −kz
Yes S is a subspace and dimS = 1
For Exercises 11-20, in the Pn space, we will use the {uk } basis extensively, where uk = tk k = 0, 1, . . . , n. It could be
shown that this forms a basis, specifically it forms an independent set, by differentiating a degree n polynomial and set it to
0, and then repeating the differentiation.
Exercise 11. f (0) = 0
n
X n
X n
X n
X
f +g = aj xj + bj xj = (aj + bj )xj ∈ S since kf = kak xk ∈ S since
j=1 j=1 j=1 j=1
(f + g)(0) = 0 kf (0) = 0
Yes S is a subspace.
n
X
f= aj xj
j=1
n
X
0= aj xj =⇒ aj = 0
j=1
dimS = n
6
Exercise 12. f 0 (0) = 0
(f + g)0 = f 0 + g 0 =⇒ (f + g)0 (0) = f 0 (0) + g 0 (0) = 0
(kf )0 = kf 0 =⇒ kf 0 (0) = 0
Yes S is a subspace.
n
X
f= aj xj
j=0
n
f 0 (0) = 0 =⇒ a1 = 0
X
f0 = jaj xj−1
j=1
n
X
f = a0 + a j xj
j=2
n
X
0 = a0 + aj xj =⇒ aj = 0, j = 0, 2, 3, . . . n
j=2
dimS = n Note that for the last step, we could’ve sited the fact that a subset of an independent set, such as the {tk } basis
for Pn , is an independent set, by definition, and so if that subset spans S, this subset will be a basis for S.
Exercise 13. f 00 (0) = 0
(f + g)00 = f 00 + g 00 =⇒ (f + g)00 (0) = f 00 (0) + g 00 (0) = 0
(kf )00 = kf 00 =⇒ kf 00 (0) = 0
Yes S is a subspace.
n
X n
X
f 00 = j(j − 1)aj xj−2 =⇒ f = a0 + a1 x + aj xj
j=2 j=3
f 00 (0) = a2 = 0
Then f is a linear combination of {1, x, x3 , x4 , . . . , xn }, dimS = n
Exercise 14. f (0) + f 0 (0) = 0
f + g + f 0 + g 0 = (f + g) + (f + g)0 =⇒ (f + g)(0) + (f + g)0 (0) = 0 + 0 = 0
kf + (kf )0 = k(f + f 0 ) =⇒ kf (0) + (kf )0 (0) = k(f (0) + f 0 (0)) = 0
Yes S is a subspace.
n
X n
X
0 j
f +f = aj x + jaj xj−1 n
X
j=0 j=1 f = a0 (1 − x) + aj xj
j=2
(f + f 0 )(0) = a0 + a1 = 0 =⇒ a0 = −a1
If f = 0, aj = 0 for j = 2, 3, . . . n, by taking j = 2, 3, . . . n derivatives. a0 = 0 for f (0) = 0. Thus {1 − x, x2 , x3 , . . . , xn }
is independent and span S and thus form a basis.
dimS = n
Exercise 15. f (0) = f (1)
(f + g)(0) = f (0) + g(0) + f (1) + g(1) = (f + g)(1)
kf (0) = kf (1)
Yes S is a subspace.
n
X
f= aj xj
j=0
n
By differentiating
X
0 0
f (0) = a0 = f (1) = a0 + aj f = (a0 ) + a2 (2x − 1) + a3 (3x2 − 1) . . .
j=1
f 00 = a2 (2) + a3 x
n n
n
X X
=⇒ aj = 0 or a1 = − aj X
f 00 = aj (jxj−2 ) = 0 if f = 0
j=1 j=2
j=2
n
X
j
=⇒ f = a0 + aj (x − x)
j=2
7
Then {xj−2 } is a subset of a basis for Pn .
aj = 0 for j = 2, . . . n.
Then a0 = 0.
Thus, {1, xj − x} is independent and spans S. Then {1, xj − x} forms a basis for S.
dimS = n
Exercise 16. f (0) = f (2)
f (0) + g(0) = (f + g)(0) = f (2) + g(2) = (f + g)(2)
kf (0) = kf (2)
Yes S is a subspace.
n
X
f= aj xj
j=0
n
X n
X n
X
f (0) = a0 = a0 + aj 2j =⇒ 2a1 + 2j aj = 0 or a1 = − 2j−1 aj
j=1 j=2 j=2
n
X
=⇒ f = a0 + aj (xj − 2j−1 x)
j=2
n
X
f 00 = aj j(j − 1)xj−2 and f 00 = 0 if f = 0
j=2
BS1 = {1, x, . . . , xn−2 } is a subset of the basis {1, x, . . . , xn } = BPn for Pn . Then BS1 is independent, and so aj = 0 for
j = 2, . . . n. Then for f = 0, a0 = 0. Thus {1, x2 − 2x, x3 − 22 x, . . . , xj − 2j−1 x, . . . , xn − 2n−1 x} is independent and
spans S and thus forms a basis for S.
dimS = n
Exercise 17. f is even. f (−x) = f (x)
(f + g)(−x) = f (−x) + g(−x) = f (x) + g(x) = (f + g)(x)
kf (−x) = kf (x)
Yes S is a subspace.
n
X n
X n
X
f (−x) = aj xj (−1)j = f (x) = aj xj =⇒ aj xj ((−1)j − 1) = 0
j=0 j=0 j=0
n
2 + 1 if n is even, is the number of possibly nonzero coefficients for f .
n−1
2 + 1 = n+1 2 if n is odd, is the number of possibly nonzero coefficients for f .
Then n2 or n−1
2 , if n is even, or n is odd, respectively, are the number of needed elements for a subset from the basis BPn to
span f and form a basis for S.
Exercise 18. f is odd. f (−x) = f (x)
(f + g)(−x) = f (−x) + g(−x) = −f (x) − g(x) = −(f + g)(x)
kf (−x) = −kf (x)
Yes S is a subspace.
n
X n
X n
X
j j j
f (−x) = aj x (−1) = −f (x) = − aj x =⇒ aj xj ((−1)j + 1) = 0
j=0 j=0 j=0
n
2 if n = 2K is even, is the number of possibly nonzero coefficients for f .
n+1
2 = n+1
2 if n = 2K − 1 is odd, is the number of possibly nonzero coefficients for f .
Then n2 or n+1
2 , if n is even, or n is odd, respectively, are the number of needed elements for a subset from the basis BPn to
span f and form a basis for S.
Exercise 19. f has degree ≤ k, where k < n or f = 0
8
k
X k
X
f= aj xj ; g= bj x j
j=0 j=0
k
X
f +g = (aj + bj )xj even if aj + bj = 0 for any or all j, f + g has degree ≤ k or f = 0
j=0
k
X
c0 aj xj ∈ S
j=0
f is spanned by BS = {xj }, j = 0, 1, . . . , k which is a subset of BPn , which is a basis for P n . Then BS is independent.
Then BS is a basis for S.
dimS = k + 1
Exercise 20. Consider
k−1
X k−1
X k−1
X
f +g = aj xj + ak xk + bj xj + −ak xk = (aj + bj )xj with ak−1 + bk−1 6= 0
j=0 j=0 j=0
f +g ∈
/S
Thus S is not a subspace.
Exercise 21.
(1) {1, t2 , t4 }
f = a0 + a2 t2 + a4 t4 , dimS = 3
(2) {t, t3 , t5 }
f = a1 t + a3 t3 + a5 t5 , dimS = 3
(3) {t, t2 }
f = a1 t + a2 t2 , dimS = 2
(4) {1 + t, (1 + t)2 }
a1 (1 + t) + a2 (1 + t)2 = a1 (1 + t) + a2 (1 + 2t + t2 ) = (a1 + a2 ) + (a1 + 2a2 )t + a2 t2
If a1 (1 + t) + a2 (1 + t)2 = 0, a2 = 0, a1 = 0 so (1 + t), (1 + t)2 is independent, dimS = 2
Exercise 22. In this exercise, L(S) denotes the subspace spanned by a subset S of a linear space V .
(1) x ∈ S, 1x ∈ L(S) =⇒ S ⊆ L(S)
P 1 , . . . , xn }
(2) If S = {x
x = aj xj ∈ L(S)
S ⊆ T =⇒ xj ∈ T P
T is a subspace of V =⇒ aj xj ∈ T (T is closed under addition and scalar multiplication). so x ∈ T
L(S) ⊆ T
(3) If S is a subspace of V , S is closed under addition and scalar multiplication.
P
Repeatedly apply addition and scalar multiplication closure for each xj ∈ S and ∀ aj ∈ R, so that aj xj ∈ S,
∀ aj ∈ R, ∀ xj ∈ S.
=⇒ L(S) ⊆ S
S ⊆ L(S) (as proven in part(a), or part (1), of this exercise).
L(S) = S if S is a subspace of V .
P
If L(S) = S, ∀ aj xj ∈ S, the S is closed under addition and scalar multiplication. Then S is a subspace of
V , by theorem.
(4) Suppose S = {x1 , x2 , . . . , xm }
Then since S ⊆ T , T = {x1 , x2 , . . . , xm , . . . xn }
m
X n
X
aj xj = aj xJ ∈ L(T ) with aj = 0 for j = m + 1, m + 2, . . . , n =⇒ L(S) ⊆ L(T )
j=1 j=1
9
\
T x1 ∈ S x2 ∈ S T
(5) If x1 , x2 ∈ S T , then \ . Since S, T are subspaces, x1 + x2 , cx1 ∈ S x1 + x2 , cx1 ∈ T
x ∈T x ∈T
T1 T2
Then x1 + x2 , cx1 T∈ S T . So S T is a subspace.
P x ∈ L(S T ). T
(6) Consider
x = aj xj ; where xj ∈ S T
Since ∀ xj ∈ S,Tthen x ∈ L(S). SinceT∀ xj ∈ T , then
T x ∈ L(T ).
Thus x ∈ L(S) L(T ). T =⇒ L(S T T ) ⊆ L(S) L(T )
(7) Example of when L(S T ) 6= L(S) L(T ).
Suppose
T S = {x1T , x2 }, T = {x3 } and
T x1 + x2 = x3 .
S T = ∅. L(S T ) = ∅, but L(S) L(T ) = {kx3 |k ∈ R} = L(T )
Exercise 23.
(1) {1, eax , ebx }, a 6= b
d
a0 + a1 eax + a2 ebx = 0 −−→ a1 aeax + a2 bebx = 0 or a1 a = −a2 be(b−a)x
dx
Since x arbitrary, a1 = a2 = 0.
=⇒ {1, eax , ebx } independent. dimS = 3
(2) {eax , xeax }
a1 eax + a2 xeax = 0 or a1 = −a2 x
x arbitrary, so a1 = a2 = 0. {eax , xeax } independent. dimS = 2
(3) {1, eax , xeax }
d
a0 + a1 eax + a2 xeax = 0 −−→ aa1 eax + a2 eax + a2 axeax = 0
dx
aa1 + a2 + a2 ax = 0 or a2 ax = −(aa1 + a2 )
x arbitrary, so a2 = 0, a1 = 0
Then a0 = 0 and so {1, eax , xeax } independent. dimS = 3
(4) {eax , xeax , x2 eax }.
a0 eax + a1 xeax + a2 x2 eax = 0 = a0 + a1 x + a2 x2
1, x, x2 are a subset of independent BPn and so 1, x, x2 are independent =⇒ a0 = a1 = a2 = 0, and so
{eax , xeax , x2 eax } independent. dimS = 3 .
(5) {ex , e−x , cosh x}
x −x
cosh x = e +e 2 dependent. dimS = 2
(6) {cos x, sin x}
a cos x + b sin x = 0 or b sin x = −a cos x
If cos x = 0, then sin x = 1, so b = 0. Otherwise,
b tan x = −a. But x arbitrary =⇒ a = 0, b = 0
So {cos x, sin x} independent. dimS = 2
(7) a cos2 x + b sin2 x = 0, so then if cos2 x 6= 0, we have b tan2 x = −a. Since x is arbitrary, a = b = 0. Then
{cos2 x, sin2 x} independent. dimS = 2
(8) {1, cos 2x, sin2 x}
cos 2x = 1 − 2 sin2 x
So the set is dependent. dimS = 2 , since {1, sin2 x} independent ({cos2 x, sin2 x} were independent and 1 =
cos2 x + sin2 x).
(9) {sin x, sin 2x}
a sin x + b sin 2x = sin x(a + b2 cos x) = 0
If sin x, cos x 6= 0, a + b2 cos x = 0 or 2b cos x = −a. Since x is arbitrary, a = b = 0 So then {sin x, sin 2x} is
independent. dimS = 2
(10) {ex cos x, e−x sin x}
aex cos x + be−x sin x = 0 or b tan x = −ae2x
Since x arbitrary, a = b = 0. dimS = 2
Exercise 24.
10
(1) Consider BS , basis for S and BV basis for V . |BV | = n finite.
If S is infinite-dimensional, then ∃ xn+1 ∈ BS s.t. xn+1 ∈ / BV since BV finite. Then ∃ xn+1 ∈ S s.t. xn+1 ∈
/ V.
But S ⊆ V
=⇒ S is finite-dimensional.
Consider BS = {x1 , . . . , xm } and BV = {y1 , . . . , yn }.
∀ xj , xj ∈ V , since S ⊆ V .
Suppose m > n. Then BS linearly dependent, by Thm. 12.10 of Vol.1 (a.k.a. Thm. 1.7 of Vol. 2). This contradicts
the fact that BS is an independent basis. =⇒ dimS ≤ dimV
(2) If S = V , then by Thm. 12.10 of Vol.1, BS must also contain exactly n vectors, since it’s a basis for V = S.
If dimS = dimV , then since BS is a set of n linearly independent elements, it forms a basis for V .
Then ∀ y ∈ V , y ∈ L(BS ) = S.
V ⊆ S =⇒ V = S.
(3) Use Thm. 12.10 of Vol.1 (a.k.a. Thm. 1.7 of Vol.2): Any set of linear independent elements is a subset of some basis
for V .
(4) Consider BV = {y1 , . . . , yn }
Suppose {y1 + y2 , y1 − y2 } = BS
y1 + y2 , y1 − y2 ∈
/ BV because if they were, they’d make BV dependent.
2.4 E XERCISES - L INEAR T RANSFORMATIONS AND M ATRICES , N ULL SPACE AND RANGE , N ULLITY AND RANK
Exercise 1. T (x, y) = (y, x)
T (a(x1 , x2 )+b(y1 , y2 )) = T (ax1 +by1 , ax2 +by2 ) = (ax2 +by2 , ax1 +by1 ) = a(x2 , x1 )+b(y2 , y1 ) = aT (x1 , x2 )+bT (y1 , y2 )
T is linear.
T (x, y) = (y, x) = 0. nullspaceT = {0}; kerT = 0
T (x, y) = (y, x) = y(1, 0) + x(0, 1). rangeT = V2 . rankT = 2
Exercise 2. T (x, y) = (x, −y)
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 , −(ax2 + by2 )) = a(x1 , −x2 ) + b(y1 , −y2 ) = aT (x1 , x2 ) + bT (y1 , y2 )
T is linear.
(x, −y) = 0. nullspaceT = {0}; kerT = 0
(x, −y) = x(1, 0) + −y(0, 1) rangeT = V2 ; rankT = 2
Exercise 3. T (x, y) = (x, 0).
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 , 0) = a(x1 , 0) + b(y1 , 0) = aT (x1 , x2 ) + bT (y1 , y2 )
T is linear.
T (x, y) = (x, 0) = 0 =⇒ x = 0, y ∈ R. nullspaceT = L({(0, 1)}) kerT = 1
T (x, y) = (x, 0) = x(1, 0) rangeT = L({(1, 0)}). rankT = 1
Exercise 4. T (x, y) = (x, x)
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 , ax1 + by1 ) = a(x1 , x1 ) + b(y1 , y1 ) = aT (x1 , x2 ) + bT (y1 , y2 )
T is linear.
T (x, y) = (x, x) = 0 =⇒ x = 0, y ∈ R. nullspaceT = L({(0, 1)}) kerT = 1
T (x, y) = (x, x) = x(1, 1) rangeT = L({(1, 1)}). rankT = 1
Exercise 5. T (x, y) = (x2 , y 2 )
T (a(x1 , x2 ) + b(y1 , y2 )) = ((ax1 + by1 )2 , (ax2 + by2 )2 ) =
= (a2 x21 , a2 x22 ) + (2abx1 y1 , 2abx2 y2 ) + (b2 y12 , b2 y22 ) 6= aT (x1 , x2 ) + bT (y1 , y2 )
T is not linear.
Exercise 6. T (x, y) = (ex , ey )
T (a(x1 , x2 ) + b(y1 , y2 )) = (eax1 +by1 , eax2 +by2 ) 6= aT (x1 , x2 ) + bT (y1 , y2 )
T is not linear.
Exercise 7. T (x, y) = (x, 1)
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 , 1) 6= a(x1 , 1) + b(y1 , 1) = aT (x1 , x2 ) + bT (y1 , y2 )
11
T is not linear.
Exercise 8. T (x, y) = (x + 1, y + 1)
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 + 1, ax2 + by2 , 1) 6= aT (x1 , x2 ) + bT (y1 , y2 )
T is not linear.
Exercise 9. T (x, y) = (x − y, x + y)
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 − ax2 − by2 , ax1 + by1 + ax2 + by2 ) =
= a(x1 − x2 , x1 + x2 ) + b(y1 − y2 , y1 + y2 ) = aT (x1 , x2 ) + bT (y1 , y2 )
T is linear.
T (x, y) = (x − y, x + y) = 0 =⇒ x = y = 0, nullspaceT = {0} kerT = 1
T (x, y) = (x − y, x + y) = x(1, 1) + y(−1, 1) rangeT = L({(1, 1), (−1, 1)}). rankT = 2
Exercise 10. T (x, y) = (2x − y, x + y)
T (a(x1 , x2 ) + b(y1 , y2 )) = (2(ax1 + by1 ) − (ax2 + by2 ), ax1 + by1 + ax2 + by2 ) =
= a(2x1 − x2 , x1 + x2 ) + b(2y1 − y2 , y1 + y2 ) = aT (x1 , x2 ) + bT (y1 , y2 )
T is linear.
(2x − y, x + y) = 0 nullspaceT = {0}. kerT = 0
x(2, 1) + y(−1, 1) = (2x − y, x + y) rangeT = L({(2, 1), (−1, 1)}). rankT = 2
Exercise 11. T rotates every point through the same angle φ about the origin. That is, T maps a point with polar coordinates
(r, θ) onto the point with polar coordinates (r, θ + φ), where φ is fixed. Also, T maps 0 onto itself.
Amazingly, T is linear . What’s required to show this is persistence.
x = (r1 cos θ1 , r1 sin θ1 )
y = (r1 , sin θ2 , r2 sin θ2 )
ax + by = (ar1 cos θ1 + br2 cos θ2 , ar1 sin θ1 + br2 sin θ2 )
|ax + by| = (ar1 c1 + br2 c2 )2 + (ar1 s1 + br2 s2 )2 =
2
= a2 r12 c21 + 2abr1 r2 c1 c2 + b2 r22 c22 + a2 r12 s21 + 2abr1 s1 r2 s2 + b2 r22 s22 =
= a2 r12 + b2 r22 + 2abr1 r2 c(θ1 − θ2 )
ar1 sθ1 + br2 sθ2
argument of ax + by = arctan
ar1 cθ1 + br2 cθ2
So |T (ax + by)| = |ax + by|, but the argument of T (ax + by) = arctan ar 1 sθ1 +br2 sθ2
ar1 cθ1 +br2 cθ2 + φ.
Consider now aT (x) + bT (y) = a(r1 , θ1 + φ) + b(r2 , θ2 + φ).
p
(ar1 c(θ1 + φ) + br2 c(θ2 + φ))2 + (ar1 s(θ1 + φ) + br2 s(θ2 + φ))2 =
p p
= (ar1 )2 + (br2 )2 + 2abr1 r2 (c(θ1 + φ)c(θ2 + φ) + s(θ1 + φ)s(θ2 + φ)) = (ar1 )2 + (br2 )2 + 2abr1 r2 c(θ1 − θ2 )
The length is the same for T (ax + by) and aT (x) + bT (y).
The argument of aT (x) + bT (y) is the following:
ar1 (sθ1 cφ + cθ1 sφ) + br2 (sθ2 cφ + cθ2 sφ) ar1 (sθ1 + cθ1 tan φ) + br2 (sθ2 + cθ2 tan φ)
= =
ar1 (cθ1 cφ − s(θ1 )sφ) + br2 (cθ2 cφ − sθ2 sφ) ar1 (cθ1 − s(θ1 ) tan φ) + br2 (cθ2 − sθ2 tan φ)
ar1 sθ1 + br2 sθ2 + ar1 cθ1 tan φ + br2 cθ2 tan φ
=
ar1 cθ1 + br2 cθ1 − ar1 sθ1 tan φ − br2 sθ2 tan φ
Beforehand, recall this trigonometric identity:
sin (x + y) sin x cos y + sin y cos x tan x + tan y
tan (x + y) = = =
cos (x + y) cos x cos y − sin x sin y 1 − tan x tan y
Thus
ar1 sθ1 +br2 sθ2
+ tan φ1
ar1 sθ1 + br2 sθ2 ar1 cθ1 +br2 cθ2
tan arctan +φ = =
ar1 cθ1 + br2 cθ2 1− ar1 sθ1 +br2 sθ2
tan φ
ar1 cθ1 +br2 cθ2
ar1 sθ1 + br2 sθ2 + ar1 cθ1 tan φ + br2 cθ2 tan φ
=
ar1 cθ1 + br2 cθ1 − ar1 sθ1 tan φ − br2 sθ2 tan φ
So the arguments for T (ax + by) and aT (x) + bT (y) are, amazingly, the same, modulo some 2π periodicity.
12
Thus, rotations are linear transformations.
nullspaceT = {0}. nullT = 0
rangeT = {(r, θ)}. rankT = 2
Exercise 12. T maps each point onto its reflection with respect to a fixed line through the origin.
We showed above that rotations are linear transformations. Then without loss of generality, consider reflection about the
x-axis.
T (a(x1 , x2 ) + b(y1 , y2 )) = (ax1 + by1 , −ax2 − by2 ) = a(x1 , −x2 ) + b(y1 , −y2 )
aT (x1 , x2 ) + bT (y1 , y2 ) = a(x1 , −x2 ) + b(y1 , y − 2)
So reflection about the x axis is linear.
Suppose R is the rotation of the fixed line into the x-axis and R is length preserving. Then for R−1 T R, reflection about any
fixed axis, (R−1 is linear too, since it’s just a rotation in the opposite direction of R)
R−1 T R(ax + by) = R−1 T (aRx + bRy) = R−1 (aT Rx + bT Ry) = aR−1 T Rx + bR−1 T Ry
T is linear.
nullT = 0 nullspaceT = {0}
rankT = 2 rangeT = {(x, y)}
Exercise 13. T maps every point onto the point (1, 1).
T (ax + by) = (1, 1) 6= aT (x) + bT (y) = (a + b)(1, 1)
T is nonlinear.
Exercise 14. T (r, θ) = (2r, θ)
T (x1 ) + T (x2 ) = 2r1 eiθ1 + 2r2 eiθ2 = 2(r1 eiθ + r2 eiθ2 )
T (x1 + y1 ) = T (r1 eiθ1 + r2 eiθ2 ) so
q
|r1 eiθ1 + r2 eiθ2 | = r12 + r22 + 2r1 r2 cos (θ1 − θ2 ) = |T (x1 ) + T (x2 )|
we see that the argument remains unchanged, while the magnitude is multiplied by 2 in each case
nullT = 0 nullspaceT = {0}
rankT = 2 rangeT = {(r, θ)}
Exercise 15. T (a(r, θ)) = (ar, 2θ) = a(r, 2θ) = aT (r, θ).
Consider this counterexample, where x1 = 1~ex , x2 = 1~ex . Not linear.
Exercise 16. T (x, y, z) = (z, y, x).
T (ax) = aT (x), T (x1 + x2 , y1 + y2 , z1 + z2 ) = (z1 , y1 , x1 ) + (z2 , y2 , x2 ) =⇒ linear
T (x) = 0 when x = 0
nullspaceT = {0} rangeT = V3
nullT = 0 rankT = 3
Exercise 17. T (x, y, z) = (x, y, 0).
T (a(x, y, z)) = a(x, y, 0) = aT (x, y, z); T (x1 + x2 , y1 + y2 , z1 + z2 ) =
= (x1 + x2 , y1 + y2 , 0) = T (x1 , y1 , z1 ) = T (x2 , y2 , z2 )
=⇒ T linear
T (x, y, z) = 0 if x = y = 0
nullspaceT = L({(0, 0, 1)}) nullT = 1
rangeT = L({(1, 0, 0), (0, 1, 0)}) rankT = 2
Exercise 18. T (x, y, z) = (x, 2y, 3z).
T (ax) = (ax, 2ay, 3az) = aT (x); T (x1 + x2 ) = (x1 + x2 , 2(y1 + y2 ), 3(z1 + z2 )) = T (x1 ) + T (x2 )
T (x) = 0 when x = y = z = 0
nullspaceT = {0} nullT = 0
rangeT = V3 rankT = 3
Exercise 19. T (x, y, z) = (x, y, 1).
T (x1 ) + T (x2 ) = (x1 + x2 , y1 + y2 , 2) 6= T (x1 + x2 ) T is not linear.
Exercise 20. T (x, y, z) = (x + 1, y + 1, z − 1)
13
T (ax + by) = (ax1 + by1 + 1, ax2 + by2 + 1, ax3 + by3 − 1) 6= aT (x) + bT (y) =
= a(x1 + 1, x2 + 1, x3 − 1) + b(y1 + 1, y2 + 1, y3 − 1)
Exercise 21. T (x, y, z) = (x + 1, y + 2, z + 3)
T (ax+by) = (ax1 +by1 +1, ax2 +by2 +2, ax3 +by3 +3) 6= aT (x)+bT (y) = a(x1 +1, x2 +2, x3 +3)+b(y1 +1, y2 +2, y3 +3)
Exercise 22. T (x, y, z) = (x, y 2 , z 3 )
T (ax + by) = (ax1 + by1 , (ax2 + by2 )2 , (ax3 + by3 )3 ) = (ax1 + by1 , a2 x22 + 2abx2 y2 + b2 y22 , a2 x23 + 2abx3 y3 + b2 y32 ) 6=
6= aT (x) + bT (y) = a(x1 , x22 , x23 ) + b(y1 , y22 , y32 )
Exercise 23. T (x, y, z) = (x + z, 0, x + y)
T (ax+by) = (ax1 +by1 +ax3 +by3 , 0, ax1 +by1 +ax2 +by2 ) = a(x1 +x3 , 0, x1 +x2 )+b(y1 +y3 , 0, y1 +y2 ) = aT (x)+bT (y)
T is linear.
x = −z
(x + z, 0, x + y) = 0 (x, y, z) = x(1, −1, −1) =⇒ nullspaceT = L({(1, −1, −1)}) kerT = 1
x = −y
(x + z, 0, x + y) = (x + z)(1, 0, 0) + (x + y)(0, 0, 1) rangeT = L({(1, 0, 0), (0, 0, 1)}), rankT = 2
Exercise 24.
n
X n
X
j
p(x) = aj x (p + q)(x) = (aj + bj )xj
j=0 j=0
n
X n
X n
X
q(x) = bj x j cp(x) = c aj xj = caj xj
j=0 j=0 j=0
n
X n
X n
X
T (p + q) = (aj + bj )(x + 1)j = aj (x + 1)j + bj (x + 1)j = T (p) + T (q)
j=0 j=0 j=0
n
X n
X
T (cp) = caj (x + 1)j = c aj (x + 1)j = cT (p)
j=0 j=0
T linear. P
n
Consider j=0 aj (x + 1)j = 0. Apply differentiation repeatedly to get aj = 0, ∀ j = 0, . . . , n.
nullspaceT = {0}. PnullT = 0.
P n j n j j
j=0 aj (x + 1) = j=0 bj x . rangeT = L({(x + 1) |j = 0, . . . , n}); rankT = n + 1
0
Exercise 25. On (−1, 1), f ∈ V ; g = T (f ), g(x) = xf (x)
T (f + g) = x(f 0 + g 0 ) = xf 0 + xg 0 = T (f ) + T (g)
T (af ) = x(af )0 = axf 0 = aT (f )
T (f ) = xf 0 (x) = 0 x is arbitrary, consider x 6= 0. f 0 (x) = 0 =⇒ f (x) = c0 .
nullspaceT = {1}, nullT = 1
rangeT = V rankT = dimV − 1 → ∞
Rb
Exercise 26. g(x) = a f (t) sin (x − t)dt for a ≤ x ≤ b
Z b Z b Z b
T (f + g) = (f (t) + g(t)) sin (x − t)dt = f (t) sin (x − t)dt + g(t) sin (x − t)dt
a a a
Z b
T (cf ) = cf (t) sin (x − t)dt
a
T is linear.
Z b Z b Z b
g(x) = f (t)(s(x)c(t) − c(x)s(t))dt = s(x) f (t)c(t)dt − c(x) f (t)s(t)dt = k1 s(x) + k2 c(x)
a a a
RangeT = L({sin x, cos x}); rankT = 2
Rb Rb
For the nullspace, now g(x) = s(x) a f (t)c(t)dt − c(x) a f (t)s(t)dt. If we take a look at Exercise 29 of this section,
then we see the answer: by the orthogonality of sin’s and cos’s, sin nt and cos nt will be orthogonal to cosRt and sinRt for
n = 2, . . . . So depending upon a and b, at least the integration over a period of 1 will result in zero for both f c and f s.
Then for the “ends” of the integration bound that don’t make a full period, make f (t) = 0. Since n = 2, · · · → ∞ for sin nt,
cos nt for the choice of f (t),
nullspaceT = L({sin nt, cos nt|n = 2, . . . }), kerT = ∞
14
Exercise 27. T (y) = y 00 + P y 0 + Qy, P, Q fixed constants.
T (ay1 + by2 ) = ay100 + by200 + P (ay10 + by20 ) + Q(ay1 + by2 ) =
a(y100 + P y10 + Qy1 ) + b(y200 + P y20 + Qy2 ) = aT (y1 ) + bT (y2 )
T is linear
nullspaceT = L({x, 1}) nullT = 2
rangeT = V rankT = ∞
Exercise 28. If x = xk is a convergent sequence with limit a, by definition,
1
∀n ∈ N, ∃m = m(n) ∈ N such that |a − xk | < ∀k ≥ m
n
T (x) = yk ; cT (x) = cyk = c(a − xk ) = ca − cxk
(cxk understood to mean that each xk term in the sequence is multiplied by c)
|c|
Consider |cxk − ca| = |c||xk − a| < for k ≥ m
n
n 1
Consider = n1 .∃m1 = m1 (n1 ) such that |cxk − ca| <
|c| n1
Thus cxk is convergent with limit ca and so T (cx) = cT (x).
Consider two convergent sequences xk and yk with limits a and b respectively. Then by definition,
1
∀n ∈ N, ∃m1 = m1 (n) ∈ N, |a − xk | < , k ≥ m1
2n
1
∀n ∈ N, ∃m2 = m2 (n) ∈ N, |b − yk | < , k ≥ m2
2n
For k ≥ max(m1 , m2 )
1 1 1
|a + b − (xk + yk )| = |(a − xk ) + (b − yk )| ≤ |a − xk | + |b − yk | < + =
2n 2n n
so when we consider T (x + y), T (x + y) = a + b − (x + y),
with a + b − (x + y) convergent sequence defined as above, with limit 0.
so T (x + y) is convergent with limit 0 just like T (x) + T (y). T is linear.
We consider a convergent sequence to be a zero if it is an additive identity to each term in the sequence. Then nullspaceT =
space of all sequences consisting of the same term for each term . Also, rangeT = space of all convergent sequences with
limit 0.
Exercise 29.
(1)
Z π Z π Z π Z π
Z π Z π
If f=
fc = f s = 0 andg= gc = gs = 0
−π −π −π −π −π −π
Z π Z π Z π Z π Z π Z π
f +g = (f + g)c = (f + g)s = 0 and kf = kf c = kf s = 0
−π −π −π −π −π −π
(by linearity of integration operation). Then S is closed under addition and scalar multiplication. S is a subspace of
V.
(2)
Z π π
1
c(nt) = s(nt) = 0
−π n −π
Z π π
−c(nt) −(c(nπ) − c(−nπ))
s(nt) = = =0
−π n
−π n
15
In general, Z π Z π
1
c(nt)c(mt) = (cos (n − m)t + cos (n + m)t)dt = 0 + 0 = 0
−π 2 −π
Z π Z π
1
s(nt)c(mt) = (sin (n + m)t + sin (n − m)t)dt = 0 + 0 = 0
−π −π 2
Z π Z π
1
s(nt)s(mt) = (cos (n − m)t − cos (n + m)t)dt = 0 + 0 = 0
−π −π 2
So S contains the functions f (x) = cos nx and f (x) = sin nx, since they satisfy the requirements.
(3) As seen above,
R the set BS = {sin nx, cos nx|n = 2, . . . } consists of orthogonal functions with an inner product
defined as over a period. Thus, they are independent of each other (orthogonal elements are independent). They
belong to S and S, being a subspace, must include all linear combinations of them, and so S is at least infinitely
dimensional, since it must contain BS in its basis.
(4) Z π Z π
T (V ) = g(x) = (1 + cos (x − t))f (t)dt = (1 + cos x cos t + sin x sin t)f (t)dt =
−π −π
Z π Z π Z π
= f (t)dt + cos tf (t)dt cos x + f (t) sin tdt sin x
−π −π −π
BT (V ) = {1, cos x, sin x}; rankT (V ) = 3
(5) T (S) = 0 =⇒ nullspaceT = S
(6) T (f ) = cf . Note that cf ∈ BT (V ) .
T (1) = 2π(1)
T (s) = πs
T (c) = πc
Exercise 30. We want the following: Let T : V → W be a linear transformation of a linear space V into a linear space W . If
V is infinite-dimensional, prove that at least one of T (V ), or N (T ), is infinite-dimensional.
Assume dimN (T ) = k, dimT (V ) = r.
Let e1 , . . . , ek be a basis for N (T ).
Let e1 , . . . , ek , ek+1 , . . . , ek+n be independent elements in V , where n > r.
Pk+n
Consider x = j=1 aj ej
k+n
X k+n
X
T (x) = aj T (ej ) = aj T (ej )(since e1 , . . . , ek ∈ N (T ))
j=1 j=k+1
x ∈ V , so T (x) ∈ T (V ). Since dimT (V ) = r, and n > r, {T (ej )|j = k + 1, . . . , k + n} must be dependent (Apostol’s
Thm.1.5 of Vol.2: any set of r + 1 elements of a dim = r space is dependent).
Then ∃ {aj }, aj ’s not all zero, s.t.
k+n
X k+n
X
aj T (ej ) = T (aj ej ) = 0
j=k+1 j=k+1
k+n
X k+n
X n
X
=⇒ aj ej ∈ N (T ) so aj eJ = aj ej
j=k+1 j=k+1 j=1
Pk+n
=⇒ j=1 aj ej = 0 is a nontrivial representation of 0. Then e1 , . . . , ek+n are dependent. Contradiction.
2.8 E XERCISES - I NTRODUCTION , M OTIVATION FOR THE CHOICE OF AXIOMS FOR A DETERMINANT FUNCTION , A
SET OF AXIOMS FOR A DETERMINANT FUNCTION , C OMPUTATION OF DETERMINANTS ,
Exercise 1. V = {0, 1}
T1 (0, 1) = 0, 0 0, 1 T1 T2 T3 T4
T2 (0, 1) = 0, 1 T1 0, 0 0, 0 1, 0 0, 0
T2 0, 0 0, 1 1, 0 1, 1
T3 (0, 1) = 1, 0 T3 1, 1 1, 0 0, 1 0, 0
T4 (0, 1) = 1, 1 T4 1, 1 1, 1 1, 1 1, 1
T2 , T3 are one-to-one, by inspection. T2−1 = T2 ; T3−1 = T3
16
Exercise 2. V = {0, 1, 2}. Note, there are obviously 33 = 27 possible ranges and thus 27 possible functions (since for each
element in V , there are 3 possible values it could be mapped to).
Consider only the 6 that are one-to-one (choice of 3 values, then 2 values, then 1 value at each subsequent stage).
T1 (0, 1, 2) = 0, 1, 2 T1−1 = T1
0, 1, 2 T1 T2 T3 T4 T5 T6
T2 (0, 1, 2) = 0, 2, 1 T1 0, 1, 2 0, 2, 1 1, 0, 2 1, 2, 0 2, 0, 1 2, 1, 0 T2−1 = T2
T3 (0, 1, 2) = 1, 0, 2 T2 0, 2, 1 0, 1, 2 2, 0, 1 2, 1, 0 1, 0, 2 1, 2, 0 T3−1 = T3
T3 1, 0, 2 1, 2, 0 0, 1, 2 0, 2, 1 2, 1, 0 2, 0, 1
T4 (0, 1, 2) = 1, 2, 0 T4 1, 2, 0 1, 0, 2 2, 1, 0 2, 0, 1 0, 1, 2 0, 2, 1 T4−1 = T5
T5 (0, 1, 2) = 2, 0, 1 T5 2, 0, 1 2, 1, 0 0, 2, 1 0, 1, 2 1, 2, 0 1, 0, 2 T5−1 = T4
T6 (0, 1, 2) = 2, 1, 0 T6 2, 1, 0 2, 0, 1 1, 2, 0 1, 0, 2 0, 2, 1 0, 1, 2
T6−1 = T6
Exercise 3. T (x, y) = (y, x) Suppose T (x1 , y1 ) = (y1 , x1 ) = T (x2 , y2 ) = (y2 , x2 ).
Then y1 = y2 , x1 = x2 → (x1 , y1 ) = (x2 , y2 )
T is one-to-one on V ; T (V2 ) = V2 , (u, v) = (y, x)
T −1 = T (by inspection).
Exercise 4. T (x, y) = (x, −y)
Suppose T (x1 , y1 ) = (x1 , −y1 ) = T (x2 , y2 ) = (x2 , −y2 )
Then x1 = x2 , −y1 = −y2 or y1 = y2 =⇒ (x1 , y1 ) = (x2 , y2 )
T is one-to-one on V , T (V2 ) = V2 , (u, v) = (x, −y)
T −1 = T
Exercise 5. T (x, y) = (x, 0).
Note that T (x, 1) = (x, 0) = T (x, 2). T is not one-to-one.
Exercise 6. T (x, y) = (x, x). Note that T (x, 1) = T (x, 2) = (x, x). T is not one-to-one.
Exercise 7. T (x, y) = (x2 , y 2 ). T (x, y) = T (x, −y) = (x2 , (−y)2 ) = (x2 , y 2 ). T is not one-to-one.
Exercise 8. T (x, y) = (ex , ey ).
Suppose T (x1 , y1 ) = T (x2 , y2 ).
Then ex1 = ex2 , ey1 = ey2 and since ex is one-to-one, ∀ x ∈ R, x1 = x2 .
T is one-to-one . u = ex , v = ey , u, v ∈ R+ . T −1 (x, y) = (ln x, ln y)
Exercise 9. T (x, y) = (x, 1) T (x, 1) = T (x, 2) = (x, 1), so T is not one-to-one.
Exercise 10. T (x, y) = (x + 1, y + 1).
If T (x1 , y1 ) = (x1 + 1, y1 + 1) = T (x2 , y2 ) = (x2 + 1, y2 + 1),
then x1 = x2 , y1 = y2 , (x1 , y1 ) = (x2 , y2 ). T is one-to-one.
u = x + 1, v = y + 1. T −1 (x, y) = (x − 1, y − 1)
Exercise 11. T (x, y) = (x − y, x + y).
If T (x1 , y1 ) = (x1 − y1 , x1 + y1 ) = T (x2 , y2 ) = (x2 − y2 , x2 + y2 )
x1 − y1 = x2 − y2
then x1 = x2 , y1 = y2 . T is one-to-one.
x1 + y1 = x2 + y2
u=x−y
T (V2 ) = L({(1, 1), (−1, 1)}); T −1 (x, y) = x+y , −x+y
v =x+y 2 2
Exercise 12. T (x, y) = (2x − y, x + y)
If T (x1 , y1 ) = (2x1 − y1 , x1 + y1 ) = T (x2 , y2 ) = (2x2 − y2 , x2 + y2 )
2x1 − y1 = 2x2 − y2 x1 = x2
so T is one-to-one.
x1 + y1 = x2 + y2 y1 = y2
u = 2x − y
T (V2 ) = L({(2, 1), (−1, 1)}). T −1 (x, y) = x+y 3 , x−2y
−3
v =x+y
Exercise 13. T (x, y, z) = (z, y, x)
If T (x1 , y1 , z1 ) = (z1 , y1 , x1 ) = T (x2 , y2 , z2 ) = (z2 , y2 , x2 )
z1 = z2
then y1 = y2 =⇒ T is one-to-one.
x1 = x2
T (V3 ) = V3 , u = z, v = y, w = z. T −1 = T
17
Exercise 14. T (x, y, z) = (x, y, 0)
T (x, y, 1) = T (x, y, 2) = (x, y, 0). T is not one-to-one.
Exercise 15. T (x, y, z) = (x, 2y, 3z)
T (x1 , y1 , z1 ) = (x1 , 2y1 , 3z1 ) = T (x2 , y2 , z2 ) = (x2 , 2y2 , 3z2 )
x1 = x2 x1 = x2
2y1 = 2y2 =⇒ y1 = y2 =⇒ T is one-to-one
3z1 = 3z2 z1 = z2
u=x
y z
v = 2y T (V3 ) = V3 T −1 (x, y, z) = x, ,
2 3
w = 3z
Exercise 16. T (x, y, z) = (x, y, x + y + z) T (x1 , y1 , z1 ) = (x1 , y1 , x1 + y1 + z1 ) = T (x2 , y2 , z2 ) = (x2 , y2 , x2 + y2 + z2 )
x1 = x2
y1 = y2 =⇒ z1 = z2 so T is one-to-one
x1 + y1 + z1 = x2 + y2 + z2
Exercise 17. T (x, y, z) = (x + 1, y + 1, z − 1)
T (x1 , y1 , z1 ) = (x1 + 1, y1 + 1, z1 − 1) = T (x2 , y2 , z2 ) = (x2 + 1, y2 + 1, z2 − 1)
x1 = x2
=⇒ y1 = y2 T is one-to-one
z1 = z2
T (V3 ) = V3 + (1, 1, −1); T −1 (x, y, z) = (x − 1, y − 1, z + 1)
Exercise 18. T (x, y, z) = (x + 1, y + 2, z + 3)
T (x1 , y1 , z1 ) = (x1 + 1, y1 + 2, z1 + 3) = T (x2 , y2 , z2 ) = (x2 + 1, y2 + 2, z2 + 3)
x1 = x2
=⇒ y1 = y2 T is one-to-one
z1 = z2
T (V3 ) = V3 + (1, 2, 3) T −1 (x, y, z) = (x − 1, y − 2, z − 3)
Exercise 19. T (x, y, z) = (x, x + y, x + y + z)
T (x1 , y1 , z1 ) = (x1 , x1 + y1 , x1 + y1 + z1 ) = T (x2 , y2 , z2 ) = (x2 , x2 + y2 , x2 + y2 + z2 )
x1 = x2
T (V3 ) = L({(1, 1, 1), (0, 1, 1), (0, 0, 1)})
y1 = y2 T is one-to-one
T −1 (x, y, z) = (x, y − x, z − y)
z1 = z2
Exercise 20. T (x, y, z) = (x + y, y + z, x + z)
T (x1 , y1 , z1 ) = (x1 + y1 , y1 + z1 , x1 + z1 ) = T (x2 , y2 , z2 ) = (x2 + y2 , y2 + z2 , x2 + z2 )
x1 + y1 = x2 + y2
x1 − z1 = x2 − z2
=⇒ y1 + z1 = y2 + z2 or =⇒ x1 = x2 , y1 = y2 , z1 = z2 so that T is one-to-one
x1 + z1 = x2 + z2
x1 + z1 = x2 + z2
T (V3 ) = L({(1, 0, 1), (1, 1, 0), (0, 1, 1)})
−1 x+z−y x+y−z y+z−x
T (x, y, z) = , ,
2 2 2
Exercise 21.
T m T n = T m T T n−1 = (T m T )T n−1 = T m+1 T n−1 = · · · = T m+n T 0 = T m+n 1 = T m+n
(T n )−1 T n = (T −1 )n T n = (T −1 )(T −1 )n−1 T (T n−1 ) = · · · =
= (T −1 ) . . . (T −1 ) T . . . (T ) = (T −1 ) . . . (T −1 )(T −1 T ) T . . . (T ) = (T −1 ) . . . (T −1 ) 1 T . . . (T ) = · · · = T −1 T = 1
| {z } | {z } | {z } | {z } | {z } | {z }
n times n times n−1 times n−1 times n−1 times n−1 times
Exercise 22.
18
(ST )n = (ST )(ST )n−1 = · · · = (ST ) . . . (ST ) =
| {z }
n times
= (ST ) . . . (ST )(ST ) = (ST ) . . . (S(T S)T ) = (ST ) . . . (ST )(SST T ) = · · · =
= SST (ST ) . . . (ST )T = · · · = S n T n
Exercise 23. (T −1 S −1 )(ST ) = T −1 S −1 ST = T −1 1T = 1, so (ST )−1 = T −1 S −1 . Its uniqueness is guaranteed by
theorem (left inverses, if they exist, are unique).
Exercise 24. (ST )−1 ST = (ST )−1 T S = 1.
Since left inverses are unique (this theorem is important to use here), then (T S)−1 = (ST )−1 =⇒ S −1 T −1 = T −1 S −1
Exercise 25.
(S + T )(S + T ) = S 2 + ST + T S + T 2 = S 2 + 2ST + T 2
(S + T )3 = (S 2 + ST + T S + T 2 )(S + T ) = S 3 + ST S + T S 2 + T 2 S + S 2 T + ST 2 + T ST + T 3 =
= S 3 + 3S 2 T + 3T 2 S + T 3
S(x, y, z) = (z, y, x)
Exercise 26.
T (x, y, z) = (x, x + y, x + y + z)
(1)
(ST ) = (x + y + z, x + y, x) T 2 = (x, 2x + y, 3x + 2y + z)
(T S) = (z, z + y, x + y + z) (ST )2 = (3x + 2y + z, 2x + 2y + z, x + y + z)
ST − T S = (x + y, x − z, −y − z) (T S)2 = (x + y + z, x + 2y + 2z, x + 2y + 3z)
S2 = 1 (ST − T S)2 = (2x + y − z, x + 2y + z, −x + 2z + y)
(2) If S(x1 , y1 , z1 ) = (z1 , y1 , x1 ) = S(x2 , y2 , z2 ) = (z2 , y2 , x2 ), then z1 = z2 , y1 = y2 , x1 = x2
If T (x1 , y1 , z1 ) = (x1 , x1 + y1 , x1 + y1 + z1 ) = T (x2 , y2 , z2 ) = (x2 , x2 + y2 , x2 + y2 + z2 ), then x1 = x2 , y1 =
y2 , z1 = z2 . Thus S, T are one-to-one.
(S −1 ) = S
(T −1 )(x, y, z) = (x, y − x, z − y)
(ST )−1 = T −1 S −1 = T −1 S T −1 S(x, y, z) = (z, y − z, x − y)
(T S)−1 = S −1 T −1 S −1 T −1 (x, y, z) = S(x, y − x, z − y) = (z − y, y − x, x)
(T − 1)(x, y, z) = (0, x, x + y)
(3) (T − 1)2 (x, y, z) = (0, 0, x)
(T − 1)3 (x, y, z) = (0, 0, 0) and for all higher powers
Rx
Exercise 27. T (p) = q(x) = 0 p(t)dt
Z x
d
DT (p) = p(t)dt = p(x) (by first fundamental thm. of calculus)
dx 0
Rx
T D(p) = 0 dtp0 (t)
Rx
Suppose p = x + 1, p0 = 1. 0 dt1 = x 6= p
nullspaceT D = {c0 | where c0 ∈ R}
rangeT D = { all polynomials p s.t. p(0) = 0 }
Exercise 28. Let V be linear space of all real polynomials p(x).
D ≡ differential operator
T is a linear map from p(x) onto xp0 (x)
(1) p(x) = 2 + 3x − x2 + 4x3
D, T, DT, T D, DT − T D, T 2 D2 − D2 T 2 .
Dp = 3 − 2x + 12x2
T p = 3x − 2x2 + 12x3 DT − T D = 3 − 2x + 12x2
DT p = 3 − 4x + 36x2 T 2 D2 − D2 T 2 = 24x − (−8 + 216x) = 8 − 192x
T Dp = −2x + 24x2
19
Pn
(2) We want T (p) = p. Try p = j=0 aj xj .
n
X
n
X n
X n
X aj (j − 1)xj = 0
T (p) = x jaj xj−1 = jaj xj = aj xj =⇒ j=0 =⇒ j = 1
j=0 j=0 j=0
aj (j − 1)xj = 0
p = a1 x
(3) We want (DT − 2D)(p) = 0 or DT (p) − 2D(p)
n
X n
X n
X n
X
DT (p) = j 2 aj xj−1 = 2 jaj xj−1 =⇒ (j 2 − 2j)aj xj−1 = j(j − 2)aj xj−1 = 0
j=0 j=0 j=0 j=0
2
j = 2, 0 so that p = a2 x + a0
(4) We want (DT − T D)n (p) = Dn (p).
n
X n
X
D aj xj = jaj xj−1
j=0 j=0
n
X n
X n
aj xj = j 2 aj xj−1
X
(DT ) (DT − T D)p = jaj xj−1 = Dp
j=0 j=0 j=0
Xn
TD = j(j − 1)aj xj−1
j=0
Dn = (DT − T D)n ∀p ∈ V
Exercise 29. xp(x). T (p) = xp.
DT (p) = D(xp) = p + xp0
=⇒ (DT − T D)(p) = p
T D(p) = T p0 = xp0
T n (p) = T n−1 (xp) = · · · = T (xn−1 p) = xn p
DT n (p) = nxn−1 p + xn p0
=⇒ (DT n − T n D)(p) = nxn−1 p = nT n−1 (p)
T n D(p) = T n (p0 ) = T n−1 (xp0 ) = · · · = T (xn−1 p0 ) = xn p0
Exercise 30.
n = 1, ST − T S = 1
n = 2, ST 2 − T 2 S = ST 2 + T (1 − ST ) = T + T = 2T
Assume the nth case, ST n − T n S = nT n−1
ST n+1 − T n+1 S = ST n T + T n (1 − ST ) = (nT n−1 )T + T n = (n + 1)T n
Pn
Exercise 31. p(x) = j=0 cj xj .
Rp = r = r(x) = p(0)
n
X
Sp = s = s(x) = ck xk−1
k=1 .
Xn
Tp = t = t(x) = ck xk+1
k=0
(1) p(x) = 2 + 3x − x2 + x3 . We want to know R, S, T, ST, T S, (T S)2 , T 2 S 2 , S 2 T 2 , T RS, RST .
T 2 S 2 = T 2 (−1 + x) = −x2 + x3
Rp = p(0) = 2 ST (p) = 2 + 3x − x2 + x3
S 2 T 2 = S 2 (2x2 + 3x3 − x4 + x5 ) = 2 + 3x − x2 + x3
Sp = 3 − x + x2 T S(p) = 3x − x2 + x3
T RSp = 3x
T p = 2x + 3x2 − x3 + x4 (T S)2 (p) = 3x − x2 + x3
RST p = 2
(2) R, S, T linear?
R(c1 p1 + c2 p2 ) = (c1 p1 + c2 p2 )(0) = c1 p1 (0) + c2 p2 (0) = c1 R(p1 ) + c2 R(p2 )
20
n1
X n2
X n2
X
c1 S(p1 ) + c2 S(p2 ) = c1 aj xj−1 + c2 bj xj−1 = hj xj−1 = S(c1 p1 + c2 p2 )
j=1 j=1 j=1
(
c1 aj + c2 bj for j = 0, . . . n1
where n2 ≥ n1 , without loss of generality, and hj =
c 2 bj for j = n1 + 1 . . . n2
n2
X
and indeed , c1 p1 + c2 p2 = hj xj
j=0
n
X n2
X n1
X n2
X n2
X
c1 p1 + c2 p2 = c1 aj xj + c2 bj xj = c1 aj xj + c 2 bj x j = hj xj
j=0 j=0 j=0 j=0 j=0
(
c1 aj + c2 bj for j = 0, . . . , n1
hj =
c2 bj for j = n1 + 1, . . . n2
n
X n1
X n2
X
T (c1 p1 + c2 p2 ) = hj xj+1 = (c1 aj + c2 bj )xj+1 + (c2 bj )xj+1
j=0 j=0 j=n1 +1
Xn1 n2
X
= c1 aj xj+1 + c2 bj xj+1 = c1 T (p1 ) + c2 T (p2 )
j=0 j=0
(3)
nullspaceR = {p| polynomial p of degree ≥ 1 }
Rp = p(0) =⇒
rangeR = {c0 |c0 ∈ R}
n
X nullspaceS = {c0 |c0 ∈ R}
Sp = ck xk−1 =⇒
rangeS = {p| polynomial p of degree n − 1 } = V
k=1
n
X nullspaceT = 0
Tp = cj xj+1
rangeT = {p| polynomial of degree ≥ 1}
j=0
(4) T is linear. nullspaceT = 0. By thm., T is one-to-one. This thm. for linear transformations is very useful because
we simply need to check if the nullspace only contains 0.
(5) If n ≥ 1, (T S)n = (1 − R)n since T S(p) = p − R(p) = (1 − R)(p).
SnT n = 1
Exercise 32. If x = {xj } is a convergent sequence, limj→∞ xj = a, let T (x) = {yn }, yn = a − xn for n ≥ 1.
V = linear space of all real convergent sequences {xj }.
T is linear, since
T (c1 x1 + c2 x2 ) = {c1 a1 + c2 a2 − (c1 x1j + c2 x2j )} = {c1 (a1 − x1j ) + c2 (a2 − x2j )} = c1 {(a1 − x1j )} + c2 {(a2 − x2j )} =
= c1 T (x1 ) + c2 T (x2 )
where limj→∞ (c1 x1j + c2 x2j ) = c1 a1 + c2 a2 .
Note that all sequences of a constant number, constant sequences, get mapped to the same sequence of zeroes. Thus, T is not
one-to-one.
2.12 E XERCISES - L INEAR TRANSFORMATIONS WITH PRESCRIBED VALUES , M ATRIX REPRESENTATIONS OF LINEAR
TRANSFORMATIONS , C ONSTRUCTION OF A MATRIX REPRESENTATION IN DIAGONAL FORM
Exercise 1.
(1) aij = δij
(2) aij = 0
(3) aij = cδij
Exercise 2.
1 0 0
(1)
0 1 0
0 1 0 0 0
(2) 0 0 1 0 0
0 0 0 1 0
0 1 0
(3)
0 0 1
21
Exercise 3.
T (3i − 4j) = −5i + 7j
(1)
T 2 (3i − 4j) = 9i − 12j
1 2 1 2 1 2 3 0
(2) T = T2 = =
1 −1 1 −1 1 −1 0 3
(3)
e1 = i − j
e2 = 3i + j
e1 + e2 e2 − 3e1 −7e1 + e2
T (e1 ) = T (i − j) = i + j − 2i + j = −i + 2j = − +2 =
4 4 4
e1 + e2 e2 − 3e1 −e1 + 7e2
T (e2 ) = T (3i + j) = 3i + 3j + 2i − j = 5i + 2j = 5 +2 =
4 4 4
1 −7 −1 1
T = , T 2 = 12
4 1 7 1
Exercise 4.
−1 0 2 1
T =2 T =4
0 1 1
Exercise 5. T : V3 → V3 be a linear transformation s.t.
T (k) = 2i + 3j + 5k
(1) T (j + k) =i =⇒ T (i + 2j + 3k) = 3i + 4j + 4k
T (i + j + k) =j−k
T (i) = −i + j − k
T (j) = −i − 3j − 5k T (c1 i + c2 j + c3 k) = (−c1 − c2 + 2c3 )i + (c1 − 3c2 + 3c3 )j + (−c1 − 5c2 + 5c3 )k = 0
T (k) = 2i + 3j + 5k
So we have a system of linear equations,
− c1 − c2 + 2c3 = 0
c1 − 3c2 + 3c3 = 0 to solve, which could be done by Gauss-Jordan or simply to add them up cleverly to get
− c1 − 5c2 + 5c3 = 0
c3 = 0 first, and then c2 , c1 = 0.
=⇒ nullspaceT
=0, nullT = 0. rangeT = V3 , rankT = 3
−1 −1 2
(2) T = 1 −3 3
−1 −5 5
Exercise 6.
e1 = (2, 3, 5) T (e1 ) = 2(−1, 1, −1) + 3(−1, −3, 5) + 5(2, 3, 5) = (5, 8, 8) = 2e1 + 2e3 + e2
e2 = (1, 0, 0) T (e2 ) = (−1, 1, −1) = −e2 + e3
e3 = (0, 1, −1) T (e3 ) = (−1, −3, −5) + (−2, −3, −5) = (−3, −6, −10) = −2e1 + 2e2
2 0 −2
T = 1 −1 1
2 1 0
Exercise 7. Given T (0) = (0, 0), T (j) = (1, 1), T (k) = (1, −1)
(1) T (4i − j + k) = (−1, −1) + (1, −1) = (0, −2)
Determine the nullspace of T by considering
c1
0 1 1 0 1 0 0 1 0
T (x) = c2 = 0 =⇒ =
0 1 −1 0 0 −2 0 0 1
c3
So then nullspaceT = L({(1, 0, 0)}).
nullT = 1
(by nullity-rank thm.)
rangeT = 2
22
0 1 1
(2) T =
0 1 −1
T (i) = 0
w1 = (1, 1)
0 1 3
(3) T (j) = w1 T =
w2 = (1, 2) 0 0 −2
T (k) = 3w1 − 2w2
(4)
e1 = i
e2 = j
3j − k
e3 =
2
T (i) = (1, 0, 1)
Exercise 8. Given ,
T (j) = (−1, 0, 1)
1 −1 2 0
Gauss-Jordan elimination c
(1) T (2i − 3j) = (2, 0, 2) + (3, 0, −3) = (5, 0, −1). 0 0 −−−−−−−−−−−−−→ 0 0 1 =⇒ c1 = c2 = 0.
c2
1 1 0 2
nullT = 0
So then
rankT = 2
1 −1
(2) Again, T = 0 0
1 1
(3)
i−j i−j w1 = (1, 0, 0)
e1 = T (e1 ) = T = (1, 0, 0)
2 2
w2 = (0, 0, 1) = k
i+j
i+j
e2 = T = (0, 0, 1) = w2 = k w3 = (0, 1, 0)
2 2
Exercise 9. Given
T (i) = (1, 0, 1)
T (j) = (1, 1, 1)
nullT = 0
(1) T (2i − 3j) = (2, 0, 2) − (3, 3, 3) = (−1, −3, −1). By inspection of the matrix for T , .
rankT = 2
1 1
(2) T = 0 1
1 1
(3) Note that
w1 = (1, 0, 1)
1 0
T (j − i) = (0, 1, 0)
w2 = (0, 1, 0) =⇒ 0 1
T (i) = (1, 0, 1) 0 0
w3 = (0, 0, 1)
Exercise 10. Let V and W be linear spaces, each with dimension 2.
T (e1 + e2 ) = 3e1 + 9e2 T (−e2 ) = −2e1 + −4e2 = −(2e1 + 4e2 )
(1) Given , then , so that T (e2 − e1 ) = e1 − e2 .
T (3e1 + 2e2 ) = 7e1 + 23e2 T (e1 ) = e1 + 5e2
nullT = 0
By inspection of matrix T ,
rankT = 2
1 2
(2) T =
5 4
(3) With a basis of (e1 , e2 ) for V and a desired basis of the form (e1 + ae2 , 2e1 + be2 ) for W ,
T (e1 ) = e1 + 5e2 =⇒ a = 5
T (e2 ) = 2e1 + 4e2 =⇒ b = 4
Exercise 11. (sin x, cos x)
0 −1 2 −1 0
D(s, c) = (c, −s) =⇒ D = D =
1 0 0 −1
Exercise 12. (1, x, ex )
23
0 1 0 1 1
D(1, x, ex ) = (0, 1, ex ) D = 0 0 0 D2 = =
0 0 1 1 1 1
Exercise 13. (1, 1 + x, 1 + x + ex )
0 1 1 1 1 1 1
D(1, 1 + x, 1 + x + ex ) D = 0 0 −1 D2 = −1 −1 = −1
0 0 1 1 1 1
Exercise 14. (ex , xex )
x x x x x 1 1 2 1 1 1 1 1 2
D(e , xe ) = (e , e + xe ) D= D = =
0 1 1 1 1
Exercise 15. (−c, s).
0 −1 −1 −1 −1
D(−c, s) = (s, c) D= D2 = =
1 0 1 1 −1
Exercise 16. (sin x, cos x, x sin x, x cos x)
D(s, c, xs, xc) = (c, −s, s + xc, c + −xs)
0 −1 1 0 −1 1 −1 1 −1 −2
1 0 0 1 1 1 1 1 −1 2
D2 =
D= =
0 0 0 −1 −1 −1 −1
0 0 1 0 1 1 −1
Exercise 17. (ex sin x, ex cos x)
1 −1 1 −1 1 −1 −2
D(ex s, ex c) = (ex s + ex c, ex c − ex s) D= D2 = =
1 1 1 1 1 1 2
Exercise 18. (e2x sin 3x, e2x cos 3x)
2 −3 2 −3 2 −3 −5 −12
D(e2x s, e2x c) = (2e2x s + 3e2x c, 2e2x c + −3e2x s) D= D2 = =
3 2 3 2 3 2 12 −5
Exercise 19. (1, x, x2 , x3 ). T (p) = xp0 .
D(1, x, x2 , x3 ) = (0, 1, 2x, 3x2 )
T (1, x, x2 , x3 ) = (0, x, 2x2 , 3x3 )
0 0 0 0
0 1 0 0
(1) T = 0 0 2 0
0 0 0 3
0 1 0 0
0 0 4 0
(2) DT = 0 0 0 9
0 0 0 0
0 0 0 0
0 0 2 0
(3) T D =
0 0 0 6
0 0 0 0
0 −1 0 0
0 0 −2 0
(4) T D − DT =
0 0 0 −3
0 0
0 0
1 1 1
(5) T 2 = =
2 2 4
3 3 9
24
(6)
2
1 6 6
T 2 D2 =
=
4
−8
9 −48
T 2 D2 − D2 T 2 =
2 8
2
2 1
6 54
D T = =
4
9
2
Exercise 20. T D = .
6
w1 = x2
3 2 2 6
Note that (T D)(x , x , x, 1) = (6x , 2x, 0, 0), so if we let , then (T D) =
w2 = x 2
2.16 E XERCISES - L INEAR SPACES OF MATRICES , I SOMORPHISM BETWEEN LINEAR TRANSFORMATIONS AND
MATRICES , M ULTIPLICATION OF MATRICES
1 2 2 2
1 −4 2
Exercise 1. A = , B = −1 3 , C = 1 −1
−1 4 −2
5 −2 1 −3
3 4
B+C =
0 2 6 −5
1 2 2 2
1 −4 2 15 −14 1 −4 2 0 0
AB = −1 3 = AC = 1 −1 =
−1 4 −2 −15 14 −1 4 −2 0 0
5 −2 1 −3
1 2 −1 4 −2 2 2 0 0 0
1 −4 2 1 −4 2
BA = −1 3 = −4 16 −8 CA = 1 −1 = 2 −8 4
−1 4 −2 −1 4 −2
5 −2 7 −28 14 1 −3 4 −16 8
30 −28 0 0 30 −28
A(2B − 3C) = 2AB − 3AC = + =
−30 28 0 0 −30 28
0 1
Exercise 2. A =
0 2
1 b11 b12 b b22
(1) AB = = 21 =0
2 b21 b22 2b21 2b22
b 0
=⇒ b21 = b22 = 0 or 11
b21 0
b11 = −2b12
b11 b12 1 b11 + 2b12
(2) BA = = =⇒
b21 b22 2 b21 + 2b22 b21 = −2b22
b11 b11 / − 2 1 −1/2 0 0
= b11 + b21
b21 b21 / − 2 0 0 1 −1/2
Exercise 3.
c=1
1 a 1
1 b 9 a=9
(1) = =⇒
1 c 6 b=6
1 d 5 d=5
(2)
a=1
1 2
c=0
a b c d
1 = 1
6 6
1 4 9 2 1 1 9 8 4 b=6
1 d = −2
25
Exercise 4. AB − BA
(1)
4 1 1 1 2 2 7 11 13
BA = −4 2 0 2 1 2 = −6 −4
1 2 2 1 2 1 1 2 3 6 6 9
A = 2 1 2
−2 9 3
1 2 3 AB = 6 10 4
4 1 1 −1 11 4
B = −4 2 0
1 2 1 −9 −2 −10
AB − BA = 6 16 8
−7 5 −5
(2)
2 0 0 3 1 −2 6 2 −4
AB = 1 1 2 3 −2 4 = 0 9 24
2 0 0 −1 2 1 −3 5 11 0 0 21
A= 1 1 2
−1 2 1 3 1 −2 2 0 0 9 −3 0
BA = 3 −2 4 1 1 2 = 0 6 0
3 1 −2 −3 5 11 −1 2 1 −12 27 21
B= 3 −2 4
−3 5 11 6 2 −4 9 −3 0 −3 5 −4
AB − BA = 0 9 24 − 0 6 0= 0 3 24
0 0 21 −12 27 21 12 −27 0
Exercise 5. An Am = Am+n .
Matrix multiplication
is associative;An Am = An−1 (AAm ) = An−1 Am+1 = · · · = A0 Am+n = Am+n
1 1 1 2
Exercise 6. A = A2 =
1 1
3 1 2 1 1 1 3
A = =
1 1 1
n 1 n n+1 n 1 1 1 n 1 n+1
Assume nth case is true A = A = AA = =
1 1 1 1
cos θ − sin θ
Exercise 7. A =
sin θ cos θ
2
c − s2
c −s c −s −2sc cos (2θ) − sin (2θ)
A2 = = =
s c s c 2sc −s2 + c2 sin (2θ) cos (2θ)
cos (nθ) − sin (nθ)
Assume nth case is true : An =
sin (nθ) cos (nθ)
cos θ − sin θ cos (nθ) − sin (nθ) cos (nθ) cos θ − sin (nθ) sin θ − cos θ sin (nθ) − sin θ cos (nθ)
An+1 = =
sin θ cos θ sin (nθ) cos (nθ) cos (nθ) sin θ + cos θ sin (nθ) − sin θ sin (nθ) + cos (nθ) cos θ
cos (n + 1)θ − sin (n + 1)θ
=
sin (n + 1)θ cos (n + 1)θ
1 1 1
Exercise 8. Let A = 0 1 1
0 0 1
26
(Assume nth case is true )
1 1 1 1 1 1 1 2 3
1 n n(n+1)
A2 = 0 1 1 0 1 1 = 0 1 2 2
An = 1 n
0 0 1 0 0 1 0 0 1
1
1 1 1 1 2 3 1 3 6 n2 +n
1 n n(n+1)
A3 = 0 1 1 0 1 2 = 0 1 3 1 1 1 2 1 n+1 2 + 2n 2
2 + 2
0 0 1 0 0 1 0 0 1 An+1 = 1 1 1 n = 1 n+1 =
1 1 1
1 1 1 1 3 6 1 4 10
(n+2)(n+1)
A4 = 0 1 1 0 1 3 = 0 1 4 1 n+1 2
0 0 1 0 0 1 0 0 1 = 1 n+1
1
1 0
Exercise 9. Given
−1 1
1 0 1 0 1 0 2 0 1
A2 = = = −
−1 1 −1 1 −2 1 −2 2 1
Consider that
A3 = 2A2 − A = 2(2A − 1) − A = 3A − 2
A4 = (2A − 1)(2A − 1) = 4A2 − 4A + 1 = 4(2A − 1) − 4A + 1 = 4A − 3
Then assume the nth case, that An = nA − (n − 1).
An+1 = nA2 − (n − 1)A = n(2A − 1) − (n − 1)A = 2nA − n − nA + A =
= (n + 1)A − n
100 1
So for n = 100, we have A =
100 1
a b
Exercise 10. If A = ,
c d
2 2
a + bc ab + bd a + bc b(a + d)
A2 = =
ac + cd bc + d2 (a + d)c bc + d2
If b = 0, d = 0, a = 0, so c = 0. So the only other way for A2 = 0 is for a = −d 6= 0. a2 + bc = 0 or a2 = −bc.
For instance,
1 ±1 1 ±1 0 0
=
∓1 −1 ∓1 −1 0 0
Exercise 11. Let
(E11 )ij = δ1i δ1j
(E12 )ij = δ1i δ2j
Aij = aij
(E21 )ij = δ2i δ1j
(E22 )ij = δ2i δ2j
(1) If AB − BA = [A, B] = 0 ∀ B ∈ M22 , then since Eij ∈ M22 , [A, Eij ] = 0 ∀ i = 1, 2, 3, 4.
P
If [A, Eij ] = 0, then since ∀ B ∈ M22 , B = bij Eij , so that
X X
[A, B] = [A, bij Eij ] = bij [A, Eij ] = 0
(2) Given [A, Eij ] = 0,
2
X 2
X
(AElm )ij = aik (Elm )kj = aik δlk δmj =
k=1 k=1
= ail δmj
=⇒ (AElm − Elm A)ij = ail δmj − δli amj = 0 or ail δmj = δli amj
2
X 2
X
(Elm A)ij = (Elm )ik akj = δli δmk akj =
k=1 k=1
= δli amj
27
if l = i, aii = ajj
if m = j, ail = δli ajj ,
if l 6= i, ail = 0
if l = i, amj = 0
if m 6= j, δli amj = 0,
if l 6= i, amj unknown
a 0
i, j is completely arbitrary, and (AElm − Elm A)ij = 0 must be true ∀ i = 1, 2, j = 1, 2, then = aI is the
0 a
A.
Exercise 12. Suppose A s.t. A2 = I.
P2
(A2 )ij = k=1 aik akj = δij
2
X
if i = j, aik aki = 1 =⇒ ai1 a1i + ai2 a2i = 1
k=1
2
X
if i 6= j, aik akj = 0 =⇒ ai1 a1j = −ai2 a2j
k=1
If a11 = 0, a12 a21 = 1 but 0 = −a12 a21 . Similar if a22 = 0. Then a11 , a22 6= 0
If a12 = 0,
a211 = 1
a222 = 1
1 0 1 0
if a21 = 0, then A = ± ,±
0 −1 0 1
if a21 6= 0, a21 a11 = −a22 a21 =⇒ a11 = −a22
If a21 = 0, a211 = a222 = 1
if a21 6= 0, then a11 a12 = −a12 a22 =⇒ a11 = −a22
If a12 , a21 6= 0, then
a211 + a12 a21 = 1a222 + a12 a21 = 1
√
1 − bc √b
=⇒
c − 1 − bc
Exercise 13. Given
AC = B
2 −1 7 6
A= , B= . Find 2 × 2 matrices C and D s.t.
−2 3 9 8 DA = B
1 3 1 7 6 1 30 26
= =C
A−1 AC = C = A−1 B 4 2 2 9 8 4 32 28
−1 1 3 1
A = −1 −1
4 2 2
DAA = D = BA 7 6 1 3 1 1 33 19
= =D
9 8 4 2 2 4 43 25
Exercise 14.
(1)
1 −1 1 0 0 −2
AB = =
0 2 1 2 2 4
AB 6= BA
1 0 1 −1 1 −1
BA = =
1 2 0 2 1 3
(2)
(A + B)2 = A2 + AB + BA + B 2
(A + B)(A − B) = A2 − AB + BA − B 2
(3) [A, B] = 0
2.20 E XERCISES - S YSTEMS OF LINEAR EQUATIONS , C OMPUTATION TECHNIQUES , I NVERSES OF SQUARE MATRICES
Exercise 1.
28
1 1 3 x 5
2 −1 4 y = 11
−1 1 z 3
1 1 3 5 1 0 4 8 1 0 0 8/5
2 −1 4 11 = 2 0 3 8 = 0 0 1 8/5
−1 1 3 −1 1 3 0 1 0 −7/5
Exercise 2. Solution doesn’t exist since
5x + 3y + 3z = 2
=⇒ x + y − z = 0 but x + y − z = 1
3x + 2y + z = 1
Exercise 3.
3 2 1 1 0 2 −8 −2 0 0 0 0
5 3 3 2 = 0 3−12 −3 = 0 1 −4 −1
7 4 5 3 1 0 3 1 1 0 3 1
x = 1 − 3z x 1 −3
=⇒ y = −1 + z 4
y = −1 + 4z
z 0 1
Exercise 4.
7 4 5 3 0 −3 12 3 1 0 3 1
= =
1 1 −1 0 1 1 −1 0 0 1 −4 −1
x + 3z = 1 x 1 −3
y = −1 + z 4
y − 4z = −1
z 0 1
Exercise 5.
x 1/5
3 −2 5 1 0 −5 14 −5 0 0 0 4 1
1 −5 0 y 14/5
1 1 −3 2 = 1 1 −3 2 = 1 1 −3 0 z = 1 + z 1
=⇒
27 2 −5 2 −5
6 1 −4 3 0 −5 14 −9 0 −5 14 0 0
u 0
Exercise 6.
1 1 −3 1 5 1 1 −3 1 5 1 1 −3 1 5 1 0 −2/3 0 1
2 −1 1 −2 2 = 0 −3 7 −4 −8 = 0 0 0 −2 8 = 0 0 0 1 −4
7 1 −7 3 3 0 −6 14 −4 −32 0 −1 7/3 −2/3 −16/3 0 1 −7/3 0 8
−2
x+ z=1 x 2/3
3 y 1
= 8 0 + z 7/3
u = −4 z 1
7 −4
y− z=8 u 0
3
Exercise 7.
x 1 0
1 1 2 3
4 0 1 1 2 3
4 0 x + y = −v y −1 −1
2 2 7 11 14 0 = 0 0 3 5 6 0 =⇒ z = 3v =⇒
z = x 0 + v 3
3 3 6 10 15 0 0 0 0 1 3 0
u = −3v u 0 −3
v 0 1
Exercise 8.
1 −2 1 2 −2 1 −2 1 2 −2 1 4 12
x+ z− u=
2
3 −1 −5 9 =
0 7 −3 −9 13
=⇒ 7 7 7
4 −1 1 −1 5 0 7 −3 −9 13 −3 9 13
y+ z− u=
5 −3 2 1 3 0 7 −3 −9 13 7 7 7
x 12/7 −1/7 4/7
y 13/7 3/7 9/7
=⇒ z = 0 + z 1 + u 0
u 0 0 1
29
Exercise 9.
1 1 2
1 1 2 2 2 x 4/3
0 −3
2 −1 3 2 = −2 if a − 8 6= 0, y = 2/3
−1
5 −1 a 6 −4 z 0
0 −6 a − 10
x −5/3 4/3
1 1 2 2 1 0 5/3 4/3
if a = 8, = =⇒ y = z −1/3 + 2/3
−3 −1 −2 1 1/3 2/3
z 1 0
Exercise 10.
(1)
x −5/7 4/7 0
y 9/7 11/7 −1
=
z 0 + z 1 + u 0
u 0 0 1
(2)
5 2 −62 −1
0 −3 −11 2 −31 0 0 −11 7/2 −19 1 0 0 −2/11 3/11
1 −1 1 −1 −2 = 0 −2 0 −1 −8 = 0 1 0 1/2 4 = 0 1 0 1/2 4
1 1 1 0 6 1 1 1 0 6 1 0 1 −1/2 2 0 0 −1 7/22 −19/11
x 3/11 2/11
y 4 −1/2
=
z 19/11 + u 7/22
u 0 1
Exercise 11.
a b d −b ad − bc −ba + ba
= = (ad − bc)I
c d −c a cd − cd −bc + ad
a b
If ad − bc 6= 0, then for A = ,
c d
1 d −b
A−1 =
ad − bc −c a
a b d −b
otherwise, if ad − bc = 0, =0
c d −c a
Use thm. from determinants.
det(AA−1 ) = detAdetA−1 = detI = 1
detA, detA−1 6= 0
2 3 4
Exercise 12. 2 1 1.
−1 1 2
2 3 4 1 0 5 8 1 2 0 1 8/5 1/5 0 2/5
2 1 1 1 = 0 3 5
1 1 = 0 0 1/5 −3/5 1 4/5 =
−1 1 2 1 −1 1 2 1 1 −1 2 −1
0 1 0 5 −8 −6 0 1 0 5 −8 −6
= 0 0 1 −3 5 4 = 0 0 1 −3 5 4 =
1 −1 0 −6 10 7 1 0 0 −3 5 4
−1 2 1
=⇒ 5 −8 −6
−3 5 4
Exercise 13.
30
1 2 2 1 1 0 2 3 −2 1 0 2 3 2 1 −5/3
2/3 4/3
2 −1 1 1 = 0 −1 −3 −6 1 4 = 0 0 −3 −7 1 5 = 1 7/3 −1/3 −5/3
1 3 2 1 0 1 0 −1 0 1 0 1 0 −1 1 1 −1 1
−5 2 4
3 3 3
−1 0 1
7 −1 −5
3 3 3
Exercise 14.
1 −2 1 1 1 0 −3 5 2 1 14
8 3
−2 5 −4 1 = 0 1 −2 2 1 = 1 8 5 2
1 −4 6 1 0 −2 5 −1 1 1 3 2 1
14 8 3
8 5 2
3 2 1
Exercise 15.
1 2 3 4 1 1 −1 −2 1 −2 1 −2 1 −2 1 −2
1 2 3 1
= 1 −1 1 −2
= 1
1 −2 1=
1 2 1 1
1 −2 1
1 −2
1 1 1 1 1 1
1 −2 1 0
1 −2 1
1 −2
1
−1
1 1/2 −1 1
2 0 2 1
3 1
=
1 −1
Exercise 16. −3
1 2 1
3 1 1/2
2 9 −3 1
2.21 M ISCELLANEOUS EXERCISES ON MATRICES
Exercise 3. Use the eigenvalue method.
1 2 x x x 1
=− =⇒ =
5 4 y y y −1
x + 2y = 6x
1 2 x x x 2
=6 =⇒ =
5 4 y y 2y = 5x y 5
Indeed, we obtain P since
" 2 √1
# "
√12 −1
#
1 2 √29 2 29
√
2
−1 = −1
5 4 √529 √
2
√30
29
√
2
" # " 2 #
−1
√ −1
√ √1
1 2 2 1 2 √29 2 6
−5 √2 −1 =
−7
√ √
29 29
5 4 √529 √
2
−1
58
" #
√2 √1
P = 29 2
√5 −1
√
29 2
P2
Exercise 4. (A2 )ij = aik akj = ail a1j + ai2 a2j = aij
k=1
If i = j, ai1 a1i + ai2 a2i = aii
it must be that i = 1 or i = 2. Then rewrite as a2ii + aij aji = aii
If i 6= j. ai1 a1j + ai2 a2j = aij
it must be that i = 1 or i = 2 and j = 2 or j = 1, respectively. then
aii aij + aij ajj = aij
31
If aij = 0, a2ii = aii . aii = 1
1 a 1 a 1 2a
= = A =⇒ a = 2a, so a = 0
0 1 0 1 0 1
If aij 6= 0, aii + ajj = 1
Note that aij , aji must be both nonzero for the following:
a2ii − aii + aij aji = 0
p
1 ± 1 − 4aij aji
aii =
2 p
1 ∓ 1 − 4aij aji
ajj = 1 − aii =
2
Exercise 5. A2 = A
(A + I)2 = 3A + I
(A + I)3 = (3A + I)(A + I) = 7A + I
(A + I)k+1 = (I + (2k − 1)A)(A + I) = A(1 + 2k − 1) + I + (2k − 1)A = (2k+1 − 1)A + I
Exercise 6.
x0 = a(x − vt)
y0 = y
z0 = z
t0 = a(t − vx/c2 )
1 −v
L(v) = a
−vc−2 1
1 + uv/c2 −u − v
1 −v 1 −u
L(v)L(u) = a b = ab −v−u
−vc−2 1 −uc−2 1 c2
uv
c2 + 1
−(u+v)
" #
c c uv 1 1+uv/c2
=√ √ 1+ 2 −(u+v)
c2 − v 2 c2 − u2 c 1+ uv 1
c2
2 !−1/2 −1/2
2uv u2 v 2 u2 + 2uv + v 2
1 −(u + v)/c uv 2
γ=q 2 = 1 − =
(1 + uv
1+ 2 + 4 −
c2
/ 1+ 2 =
1 − wc c2 ) c c c
−1/2
u2 v 2 u2 v2
uv 2
= 1+ 2 − 2 − 2 / 1+ 2
c c c c
Exercise 7.
(1) (AT )ij = Aji
(AT )Tij = (AT )ji = Aij =⇒ (AT )T = A
(2) (A + B)Tij = (A + B)ji = Aji + Bji = ATij + Bij
T
=⇒ (A + B)T = AT + B T
T T
(3) (cA)ij = (cA)ji = cA ji = c(A )ij
(4) (AB)Tij = (AB)ji = k ajk bki = k bki ajk = k (B T )ik (AT )kj = (B T AT )ij
P P P
(5) A−1 A = 1 =⇒ (A−1 A)T = AT (A−1 )T = 1 then (A−1 )T = (AT )−1 (recall that a right inverse is also a left
inverse).
cos θ − sin θ cos θ sin θ 1 0
Exercise 8. =
sin θ cos θ − sin θ cos θ 0 1
n
X n
X
Ai Aj = aik ajk = aik aTkj = (AAT )ij = δij
k=1 k=1
Exercise 9.
AAT = 1
(1) If , (A + B)(A + B)T = 2 + BAT + AB T
BB T = 1
(2) (AB)(AB)T = (AB)B T AT = 1
(3) B is given to be orthogonal.
32
Exercise 10.
(1)
1 1 1 1 1 −1 1 −1 −1 1 −1 1 −1 −1 −1 −1
1 −1 −1 1 1 1 1 −1 1 1 −1 −1 1 −1 −1 −1
(2) (X + Y ) · (X + Z) = X 2 + X · Z + Y · X + Y · Z = X 2 Lemma 1 is true.
Lemma 2 (xi + yi )(xi + zi ) = x2i + xi (zi + yi ) + yi zi
xi yi zi
1 1 1 4
1 1 −1 0
1 −1 −1 0
1 −1 1 0
−1 1 1 0
−1 −1 1 0
−1 1 −1 0
−1 −1 −1 4
Assume A is Hadamard.
Then by Lemma 1, (Ai + Aj ) · (Ai + Ak ) = A2i = n, i, j, k distinct.
n
X n
X n
X n
X
(Ai + Aj ) · (Ai + Ak ) = A2i + Ai · Ak + Aj · Ai + Aj · Ak = a2il + ail akl + ajl ail + ajl akl =
l=1 l=1 l=1 l=1
n
X
= (ail + ajl )(ail + akl )
l=1
By Lemma 2, (ail + ajl)(ail +Pakl ) = 0 or 4
n
then (Ai + Aj ) · (Ai + Ak ) = l=1 (ail + ajl )(ail + akl ) = 4m, where m ≤ n
=⇒ n = 4m
3.6 E XERCISES - I NTRODUCTION , M OTIVATION FOR THE CHOICE OF AXIOMS FOR A DETERMINANT FUNCTION , A
SET OF AXIOMS FOR A DETERMINANT FUNCTION , C OMPUTATION OF DETERMINANTS ,
Exercise 1.
2 1 1
(1) 1 4 −4
1 0 2
0
1 −3 0 1 −3 1 0 2
0
4 −6 = 0 0 −6 = 0 1 −3 = 6
1 0 2 1 0 2 0 0 −6
3 0 8
(2) 5 0 7 = 3(−28) + 8(20) = 4(−3(7) + 2(20)) = 76
−1 4 2
a 1 0
(3) 2 a 2 = a(a2 − 2) − 1(2a) = a(a2 − 4) = a(a − 2)(a + 2)
0 1 a
x y z
Exercise 2. Given det 3 0 2 = 1,
1 1 1
2x 2y 2z
(1) 32 0 1
1 1 1
2x 2y 2z
3 1
2 0 1 =2
detA = 1
1 2
1 1
33
x y z
(2) 3x + 3 3y
3z + 2
x+1 y+1
z+1
x y z x y z x y z x y z x y z
3x + 3 3y 3z + 2 = 3x + 3 3y 3z + 2 + 3x + 3 3y 3z + 2 = 0 + 3x 3y 3z + 3 0 2 = 1
x+1 y+1 z+1 x y z 1 1 1 1 1 1 1 1 1
x−1 y−1 z−1
(3) 4 1 3
1 1 1
x − 1 y − 1 z − 1 x y z −1 −1 −1 x y z
4 1 3 = 4 1 3 + 4 1 3 = 3 + 1 0 + 1 2 + 1 = 1
1 1 1 1 1 1 1 1 1 1 1 1
Exercise 3.
(1)
1 1 1 1 1 1
0 b − a c − a = 0 b−a c−a = (b − a)((c − a)(c + a) − (c − a)(b + a)) =
0 b2 − a2 c − a2 0
2
0 2 2
(c − a ) − (c − a)(b + a)
= (b − a)(c − a)(c − b)
(2)
1 1 1 1 1 1 1 0 0
a
3 b c = 0 b−a c − a = 0 (b − a) c−a
a b3 c3 0 b3 − a3 c3 − a3 0 (b − a)(b2 + ba + a2 ) 2 2
(c − a)(c + ca + a )
subtract the second column off the third column modulo a factor
0−0
c − a − (c − a)
2 2 2 2
(c − a)(c + ca + a − (b + ba + a ))
1 0 0 1 0 0
= 0 (b − a) 0
= 0 (b − a) 0
0 (b − a)(b2 + ba + a2 ) (c − a)(c2 − b2 + ca − ab) 0 0 (c − a)(c − b)(c + b + a)
= (a + b + c)(c − a)(c − b)(b − a)
1 1 1 1 1 1 1 0 0
2 2 2
a b c = 0 b2 − a2 c2 − a2 = 0 (b − a)(b + a) (c − a)(c + a)
3 3 3
a b c 0 b3 − a3 c3 − a3 0 (b − a)(b2 + ba + a2 ) (c − a)(c2 + ca + a2 )
subtract the second column off the third column modulo a factor
0 0
(c − a)(c + a)
(c − a)(c + a) − (b − a)(b + a) =
2 2 (b − a)(b + a) 2 2
(c − a)(c + ca + a ) (b − a)(b + ba + a )
0 0
= 0 = 0
2 2 (c+a) 2 2 (c−a)
(c − a)(c + ac + a − (b+a) (b + ba + a )) (b+a) (c − b)(ac + ab + bc)
= (b − a)(c − a)(c − b)(ac + ab + bc)
Exercise 4.
(1)
1
−1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
1 −1 −1 −1 −1 −1 1 1 0 0 2 2 0 0 2 2
= = = = (−1)8
1
1 −1 −1 1 −1 −1 1 2 0 0 2 0 0 0 2
1 1 1 −1 1 −1 −1 −1 2 0 0 0 2 0 0 0
34
(2)
1 1 1 1 1 1 1 1 1 0 0 0
a b c d 0 b−a c−a d − a 0 b−a b2 − a2 b3 − a3
2 = =
a
3 b2 c2 d2 0 b2 − a2 c2 − a 2 d2 − a2 0 c−a c2 − a2 c3 − a2
a b3 c3 d3 0 b3 − a3 c3 − a 3 d3 − a3 0 d−a d2 − a2 d3 − a3
1 0 0 0
2 2
0 1 (b + a) b + ba + a
= (b − a)(c − a)(d − a) =
0 1 (c + a) c2 + ac + a2
0 1 (d + a) d2 + ad + a2
0
1
(Now I use the addition of column , which doesn’t change the determinant)
1
1
1 0 0 0
0 1 b (b + a)b
= (b − a)(c − a)(d − a) =
0 1 c (c + a)c
0 1 d (d + a)d
1 0 0 0
0 1 0 0
= (b − a)(c − a)(d − a)
0 0 c − b (c + a)c − (b + a)b
0 0 d − b (d + a)d − (b + a)b
1 0 0 0
0 1 0 0
(b − a)(c − a)(d − a) =
0 0 c−bc − b2
2
2 2
0 0 d−bd −b
1 0 0 0
0 1 0 0
= (b − a)(c − a)(d − a)(c − b)(d − b) =
0 0 1 c + b
0 0 1 d + b
(b − a)(c − a)(d − a)(c − b)(d − b)(d − c)
(3)
(4)
a 1 0 0 0 a 1 a 0
4
4
4
a 2 0 0 a− a 2
a− a 2
0 3 a 3 0 = 3 a 3 = 3 a 3
4
0
0 2 a 4 2 a 4 2 a− a 4
0 0 0 1 a 1 a 0 a
a
4
a−
a
12
a−
= 0 0 4
a− a
=
0a − a4 0
0 a
4 12 4
= a2 (a − )2 a − 4 = a2 (a − )(a2 − 4 − 12)
a a− a a
= a(a2 − 4)(a2 − 16)
35
(5)
1 1 1 1 1 1 1 1
1
1 1 −1 −1 −1 0 0 0 −2 −2 −2 2 2 −2 −2 0
1 1 −1 −1 1 1 −2 −2 −2 −2
= =
1
−1 −1 1 −1 1 −2 −2 −2
−2 −2 −2
1
−1 1 −1 1 1 −2 −2
−2 −2
1 −1 −1 1 1 −1 −2 −2 −2 0 0 −2
1
2
−2 −2
= (1)(2)(−2)(−2)(10)(−2) = −160
=
−2 −4
0 −6 −2
0 0 −2
Exercise 5. Consider A = (aij ) s.t. aij = 0 whenever i < j.
Suppose a11 = 0. Then a1j = 0, ∀ j ≤ n, since a row of A is entirely zero, by homogeneity property of determinants,
detA = 0.
Suppose aii = 0 for some 1 < i ≤ n.
then i rows have n − (i − 1) components equal to zero. Therefore, these i rows can span a psace of at most i − 1 dimensions.
then the i rows are dependent. Then detA = 0 by Thm. (the determinant vanishes if its rows are dependent).
Then assume aii nonzero ∀ i ≤ n
Let An = Bn + Cn , where Bn = 0 0 ... 0 ann and Cn = an1 an2 ... an, n−1 0
det(A) = det(A1 , A2 , . . . , An ) = det(A1 , A2 , . . . , Bn + Cn ) = det(A1 , A2 , . . . , Bn ) + det(A1 , A2 , . . . , Cn ) =
= det(A1 , A2 , . . . , Bn )
Also, An−1 = Bn−1 +Cn−1 , where Bn−1 = 0 0 ... an−1,n−1 0 and Cn−1 = an−1,1 an−1,2 ... an−1, n−2 0 0
det(A) = det(A1 , A2 , . . . , An−1 , Bn ) = det(A1 , A2 , . . . , Bn−1 + Cn−1 , Bn ) =
= det(A1 , A2 , . . . , Bn−1 , Bn ) + det(A1 , A2 , . . . , Cn−1 , Bn ) =
= det(A1 , A2 , . . . , Bn−1 , Bn )
Then detA = det(B1 , B2 , . . . , Bn ). Qn
By homogeneity of determinants, detA = i = 1n aii detI = i=1 aii
Q
Exercise 6.
F = f1 g2 − f2 g1
F 0 = f10 g2 + f1 g20 − f20 g1 − f2 g10 = f10 g2 − f20 g1 + f1 g20 − f2 g10 =
0
f1 f20 f1 f2
= +
g1 g2 g10 g20
Exercise 7.
g2 g3 g1 g3 g g2
F = f1 − f2
+ f3 1
h2 h3 h1 h3 h1 h2
0 0 0
f1 f2 f3 0
g30 g2
0
g30 g1
0
g20 g1
g g3 g1 g3 g1 g2
F 0 = g1 g2 g3 + f1 2
+ − f 2 + + f 3 + =
h2 h3 h02 0
h3 h01 0
h2 h01 h02
h1 h2 h3 h3 h1 h3 h1
0
f1 f20 f30 f1 f2 f3 f1
f2 f3
= g1 g2 g3 + g1 g20 g30 + g1
0
g2 g3
h1 h2 h3 h1 h2 h3 h01 h02 h03
Exercise 8. Using the previous results:
(1)
0
f20 f1
0
f g1 f1 g1
F = 10 + =
f20 f200 g200
00
f1 f2 g200
36
(2)
f1 f2 f3 f1 f2 f3
F = f10 f20 f3 ] =⇒ f10
0
f20 f30
f100 f200 f300 f100 f200 f300
Exercise 9.
(1) ( ( (
uij if i < j vij if i < j uij + vij if i < j
(U + V )ij = uij + vij = + =
0 otherwise 0 otherwise 0 otherwise
n
(P
X X X
i<k,k<j uik vkj if i ≤ j
(U V )ij = uik vkj uik vkj = uik vkj =
k=1 i<k i<k,k<j
0 otherwise
(2)
n n n n
! n
!
Y X Y Y Y
det(U V ) = uik vki = uii vii = uii vii = detU detV
i=1 i<k,k<j i=1 i=1 i=1
(3) Suppose U U −1 = 1.
xdet1 = 1 = (detU )(detU −1 ), U and 1 are 2 n × n triangular matrices
U −1 exists since detU 6= 0.
detU −1 = 1/detU
(cof A)T
Exercise 10. Use the cofactor matrix to get the inverse, that detA = A−1 .
1
detA = 16, detA−1 =
16
−3
1 1 1
2 4 8 16 2 3 4 5
0 1 −3 1
2 4 8 0 2 3 4
−3
1
0 0 2 4 0 0 2 3
1
0 0 0 2
0 0 0 2
3.11 E XERCISES - T HE PRODUCT FORMULA FOR DETERMINANTS , T HE DETERMINANT OF THE INVERSE OF A
NONSINGULAR MATRIX , D ETERMINANTS AND INDEPENDENCE OF VECTORS , T HE DETERMINANT OF A
BLOCK - DIAGONAL MATRIX
Exercise 1.
−3 5 detA = −1
1 3 −4 2
(1) If A = ;B = ;A+B = det(A + B) = −58
2 5 3 6 5 11 detB = −30
(2) det(A + B)2 =det(A + B)(A +B) = det(A+ B)det(A
+ B) = (det(A + B))2
1 1 −1 2 0 3
(3) If A = ;B = ,A+B =
−1 1 1 −1 0 0
1 1 1 1 0 2
A2 = =
−1 1 −1 1 −2 0
2 2 0 2 0 1 3 −4 0 2 0 2
A + 2AB + B = +2 + = +
−2 0 2 −3 −2 3 −2 0 4 −
−1 2 −1 2 3 −4
B2 = =
1 −2 1 −2 −2 3
=⇒ det(A2 + 2AB + B 2 ) = −9
3 −2
(4) likewise, det(A2 + B 2 ) = det =1
−4 3
Exercise 2.
(1) Assume
A is n × n, B is m × m, and C is p × p.
A 0
is a n + m × n + m matrix, D.
0 B
A
D D 0
B = . Then by Thm. 3.7, det = detDdetC
C 0 C
C
37
A
A 0
detD = det = detAdetB (by Thm. 3.7). Then det B = detAdetBdetC
0 B
C
A1
A2 Q
n
(2) Assume det = i=1 detAi .
. .
.
An
A1
A2
Consider det
. ..
An+1
A1
A2 Pn Pn
Now = Dn , a square matrix of size i=1 Ni × i=1 Ni where Ni = size of matrix Ai .
..
.
An
D Qn
det n = detDn detAn+1 by Thm. 3.7. detDn , by induction assumption, is detDn = i=1 detAi .
A
n+1
A1
A2 Q
n+1
=⇒ det = i=1 detAi
..
.
An+1
Exercise 3.
1 1
1 1 = det c
d
detA = deta b
= det
c d c d g h
e f g h g h
a b c d a b
e f g h e f = det a b
detB = det0 0 =
1 0 0 0 1 0 e f
0 0 0 1 0 0 0 1
A
Exercise 4. If X = where A is (n − m) × n and I is m × m, then ∀ aij entry, (n − m) + 1 ≤ i ≤ n,
Im
(n − m) + 1 ≤ j ≤ n.
[0 0 . . . 0, −aij , 0, . . . 0] could be added to the ith row since Gauss-Jordan row operations do not change the determinant,
An−m
by determinant properties. Then detX = . By Thm. 3.7, detX = detAn−m .
Im
I
Similarly for Y = m .
A
a b 0 0
c d 0 0 a b g h
Exercise 5. A = detA = det det
e f g h c d z w
x y z w
B 0
Exercise 6. A = where B is m × m, C, D are (n − m) × (n − m).
C D
detA = f (A1 , A2 , . . . , An ), Ai = Ci + Di , m+1≤i≤n
detA = f (A1 , A2 , . . . , An ) = f (A1 , A2 , . . . , Cm+1 + Dm+1 , . . . , An ) =
= f (A1 , A2 , . . . , Cm+1 , . . . , An ) + f (A1 , A2 , . . . , Dm+1 , . . . , An )
Consider A1 , A2 , . . . , Cm+1 , m + 1 rows with m possibly nonzero components. Then A1 , . . . , Cm+1 span at most a dimm
subspace. Then A1 , . . . , Cm+1 dependent. By Thm., f (A1 , A2 , . . . , Cm+1 , . . . , An ) = 0
detA = f (A1 , A2 , . . . , Dm+1 , . . . , An )
38
Likewise for i = m + 2, . . . , n
=⇒ detA = f (A1 , A2 , . . . , Dm+1 , . . . , Dn ) = detBdetD
(By Thm. for det of block-diagonal matrices)
Exercise 7.
1 −1 0 1 −1 0 1 −1 0
(1) 0 1 −1 = 0 1 −1 = 0 1 −1 = 4
2 3 −1 −1 0 0 4
0 5
1 −1 2 1 0 −1 2 0 0 −1 2 0 0 −1 2 0 1 0 0 1
−1 2 −1 0 0 2 −1 1 0 0 3 1 0 0 0 −8 0 −1 0 0
(2)
= = = = =
3 −1 1 0 0 −1 1 −3 0 0 −1 −3 0 0 −1 −3 0 0 −1 −3
1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 0 0 0 −8
−8
1 0 0 01 0 0 0 0 1 0 0 0 0 1
1 1 0 0 0 1 0 0 0 0 1 0 0 0 0
(3) 1 0 1 0 1 = 0 0 1 0 1 = 0 0 1 0 0 = 0
1 1 0 1 1 0 0 0 0 1 0 0 0 0 0
0 1 0 1 0 0 1 0 1 0 0 1 0 1 0
3.17 E XERCISES - E XPANSION FORMULAS FOR DETERMINANTS . M INORS AND COFACTORS . 3.13 E XISTENCE OF THE
DETERMINANT FUNCTION , T HE DETERMINANT OF A TRANSPOSE , T HE COFACTOR MATRIX , C RAMER ’ S RULE
Exercise 1.
1 2 4 −3
(1) cof A =
3 4 −2 1
2 −1 3 2 −1 1
(2) 0 1 1 cof A = −6 3 5
−1 −2 0 −4 −2 2
3 1 2 4 109 113 −41 −13
2 0 5 1 −40 −92 74 16
(3) cof A =
1 −1 −2 6 −41 −79 7 −47
−2 3 2 3 −50 38 16 20
Exercise 2.
1 4 −3
(1) −2 −2 1
2 −6 −4
(2) 81 −1 3 −2
1 5 2
109 −40 −41 −50
1 113
−92 79 38
(3) 184 −41 74
7 16
−13 16 −47 20
Exercise 3. Note that for λI − A, det(λI − A) = 0 = det(A − λI)
(1)
−λ 3
= 0 =⇒ λ + λ2 − 6 = 0 = (λ + 3)(λ − 2) = 0
2 −1 − λ
(2)
1 0 2
0 −1 −2
2 −2 0
1 − λ 2
−1 − λ −2 = (1 − λ)(λ(1 + λ) − 4) + 2(2(1 + λ)) =
2 −2 −λ
= (1 − λ)(λ2 + λ − 4) + 4(1 + λ) = (1 − λ)(λ2 + λ − 4) + 4(1 + λ) =
= −λ3 + 9λ = λ(−λ2 + 9) =⇒ λ = ±3, 0
39
(3)
11 − λ −2 8 3 − λ 0 3 − λ 0 0 3 − λ
19
−3 − λ 14 = 19 −3 − λ 14 = 5 −3 − λ 14
−8 2 −5 − λ −8 2 −5 − λ λ − 3 2 −5 − λ
= (3 − λ)(10 + λ2 − 9) = (3 − λ)(1 + λ2 )
λ = 3, ±i
Exercise 4.
(1) ((cof A)T )ij = (cof A)ji = (−1)i+j detAji = (−1)i+j det(Aji )T = (−1)i+j det(AT )ij = cof (AT )ij
P use A(cof
(2) See Part (c), and then A)T = P(detA)I, Thm. 3.12.
T T
(3) ((cof A) A)ij = k (cof A)ik akj = k akj (cof A)ki
Recall that column expansions can be done on determinants, and that detA = detAT .
Consider B matrix whose jth column is equal to the ith column for some j 6= i,
but remaining rows are the same as A.
then detB = 0
Xn
detB = bkj cof bkj (jth column expansion of B)
k
bkj = aij
cof bkj = cof akj (since B differs from A only in the jth column)
n
X
=⇒ aij cof akj = 0
k
P
If i = j, k aki (cof A)ki = detA (by ith column expansion of detA)
Exercise 5.
x + 2y + 3z = 8
1 2 3 x
8
(1) 2x − y + 4z = 7 =⇒ 2 −1 4 y = 7
−y + z = 1 −1 1 z 1
8 2 3 0 10 −5 1 −1 1
1 1 1
x= 7 −1 4 = 0 6 −3 = 0 10 −5 = 0
−7 −7 −7
1 −1 1 1 −1 1 0 6 −3
1 8 3 1 0 −5
1 1
2 0 −3 = −7/ − 7 = 1
y= 2 7 4 =
−7 −7
1 1 1 0
1 2 8
−1 1 0 10
z=
1
2 −1 7 = 0 −1 −13 = −14 = 2
−7 7 −7
−1 1 −1 1
x + y + 2z = 0
1 1 2
x 0
(2) 3x − y − z = 3 =⇒ 3 −1 −1 y = 3
2x + 5y + 3z = 4 2 5 3 z 4
1 1 2 1 1 2
3 −1 −1 = 0 −4 −7 = 25
2 5 3 0 3 −1
0 1 2 0 1 2
3 −1 −1 = 3 0 1 = −(−21 − 4) = 25 =⇒ x = 1
4 5 3 4 0 −7
1 0 2 1 0 2
3 3 −1 = 0 3 −7 = 25 =⇒ y = 1
2 4 3 0 4 −1
z = −1
Exercise 6.
40
(1) Vector form of lines: tA + P1 = X; A = P2 − P1 .
x2 − x1 x1 x
t + =
y2 − y1 y1 y
x − x1 x − x1
t 2 +− =0
y2 − y1 y − y1
Then A, X − P1 are linearly dependent.
Then if A, X − P1 form rows of a matrix,
x − x1 y − y1
x2 − x1 =0
y2 − y1
Also
x x x
t 2 + (1 − t) 1 − =0
t2 t1 y
we can extend this to say
x2 x1 x1
t y2 + (1 − t) y1 − y1 = 0 =
1 1 1
= tX2 + (1 − t)X1 − X = 0
X, X1 , X2 are dependent, and so if X, X1 , X2 form rows of a matrix, then
x y 1
x1 y1 1 = 0
x2 y2 1
(2) Recall the vector form for planes: P = {X|X = P + sA + tB}, A, B are independent.
x x0 x1 − x0 x2 − x0
X = y = y0 + s y1 − y0 + t y2 − y0 = P + sA + tB
z z0 z1 − z0 z2 − z0
0 = P − X + sA + tB
P − X, A, B are dependent. Then consider P − X, A, B to be rows of a matrix. Then
x0 − x y0 − y z0 − z
x1 − x0 y1 − y0 z1 − z0 = 0
x2 − x0 y2 − y0 z2 − z0
We could also rewrite this equation like this:
x x0 x1 x0 x2 x0
y − y0 − s y1 + s y0 − t y2 + t y0 = 0
z z0 z1 z0 z2 z0
x x0 x2 x1
y + (t + s − 1) y0 − t y2 − s y1 = 0
z z0 z2 z1
Extend by 1 for a new row.
x x0 x2 x1
y
+ (t + s − 1) y0 − t y2 − s y1 = 0
z z0 z2 z1
1 1 1 1
This shows that these 4 vectors are linearly dependent. Consider the vectors as rows of matrix to obtain:
x y z 1
x0 y0 z0 1
x1 y1 z1 1 = 0
x2 y2 z2 1
41
(3) We have 3 noncollinear points that satisfy some specific equation for a circle in the x − y plane.
(x − x0 )2 + (y − y0 )2 = ρ2 = (x2 − 2xx0 + x20 ) + (y 2 − 2yy0 + y02 ) or x2 − 2x0 x + y 2 − 2y0 y − (ρ2 − x20 − y02 ) = 0
So x0 , y0 , the coordinates for the origin, and ρ, the radius of the circle, are 3 unknowns and 3 equations are needed.
To fix the “scale” of the coordinates, we need a 4th equation.
x21 − 2x0 x1 + y12 − 2y0 y1 − (ρ2 − x20 − y02 ) = 0 2
x + y2
x y 1
x22 − 2x0 x2 + y22 − 2y0 y2 − (ρ − 2
x20 − y02 ) =0 x21 + y12
−2x0 −2y0 −(ρ2 −x20 −y02 ) 1 = 0
x1 y1
=⇒ =⇒ 2 2
x2 + y2 x2 y2 1
x23 − 2x0 x3 + y32 − 2y0 y3 − (ρ2 − x20 − y02 ) =0
2 2 2 x23 + y32 x3 y3 1
x − 2x0 x + y − 2y0 y − (ρ − x20 − y02 ) =0
Notice how x2j and yj2 must be “correlated” in that their relative values are not independent, but must be 1 to 1.
We can also consider “getting rid” of the ρ2 unknown by taking an equation minus the previous equation:
x21 − 2x1 x0 + x20 + y12 − 2y1 y0 + y02 = ρ2
−(x2 − 2xx0 + x20 + y 2 − 2yy0 + y02 = ρ2 )
=⇒ x21 − x2 + −2x0 (x1 − x) + y12 − y 2 + −2y0 (y1 − y) = 0
So that we get
x21 − x2
2
y1 − y 2
x1 − x y1 − y
x22 − x21 2 2
+ −2x0 x2 − x1 + y22 − y12 + −2y0 y2 − y1 = 0
2 2
x3 − x2 x3 − x2 y3 − y2 y3 − y2
x2 − x23 x − x3 y 2 − y32 y − y3
Thus, these 4 vectors above are linearly dependent, which implies
2
x1 − x2 x22 − x21 x23 − x22 x2 − x23
x1 − x x 2 − x1 x3 − x2 x − x3
2
y1 − y 2 y22 − y12 y32 − y22 =0
y 2 − y32
y1 − y y2 − y1 y3 − y2 y − y3
Exercise 7. F (x) = det[fij (x)]
i=1 f11 =⇒ |f11 | = F (x) =⇒ F 0 (x) = f11 0
= detA1
f11 f12
i=2 f21 f22 = f11 f22 − f12 f21 = F (x)
F 0 (x) = f11
0
f22 + f11 f220 0
− f12 0
f21 − f12 f21
0 0
f11 f12 + f11 f12 0 0 0 0 0
0 0 = |A1 | + |A2 | = f11 f22 − f21 f12 + f11 f22 − f12 f21 = F (x)
f21 f22 f21 f22
Assume n case is true.
n+1
X
F (x) = det(fij (x)) = fn+1, k (−1)n+1+k det(f )n+1, k
k=1
n+1
X n+1
X
F 0 (x) = 0
fn+1,k cof (f )n+1,k + fn+1,k (−1)n+1+k (det(f )n+1,k )0
k=1 k=1
Pn+1 0
k=1 fn+1,k cof (f )
Pn+1,k
= detAn+1 , matrix obtained by differentiating the n + 1 row of [fij ]
n
(det(f )n+1,k )0 = l=1 detBl where Bl is the matrix obtained by differentiating the lth row of (f )n+1,k , l = 1, . . . , n
n+1
X n+1
X n
X n n+1
X X
fn+1,k (−1)n+1+k (det(f )n+1,k )0 = fn+1,k (−1)n+1+k detBl = fn+1,k (−1)n+k+1 detBl =
k=1 k=1 l=1 l=1 k=1
n
X
= detAl
l=1
n
X n+1
X
=⇒ F 0 (x) = detAn+1 + detAl = detAl
l=1 l=1
42
(i−1)
Exercise 8. Consider W (x) = [uj (x)].
(i−1) Pn
|W (x)| = |[uj Use Ex.7: F 0 (x) =
(x)]| i=1 detAi (x), where Ai (x) is the matrix obtained by differentiating the func-
tions in the ith row of [fij (x)], then
Pn (i−1)
|W (x)|0 = i=1 detAi (x), where Ai (x) is the matrix obtained by differentiating the functions in the ith row of [uj (x)].
For i = 1, . . . , n − 1, i + 1 = 2, . . . , n and there’s a k = i + 1 row s.t. k = 2, . . . , n so that detAi = 0,
(n−1) (n) (k−1) (n)
For i = n, [uj (x)]0 = [uj (x)] and for k = 1, . . . , n − 1, [uj (x)] is different from [uj (x)].
(i−1)
=⇒ |W (x)|0 = detAn (x), where An (x) is the matrix obtained by differentiating the functions in the nth row of [uj (x)]
4.4 E XERCISES - L INEAR TRANSFORMATIONS WITH DIAGONAL MATRIX REPRESENTATIONS , E IGENVECTORS AND
EIGENVALUES OF A LINEAR TRANSFORMATIONS , L INEAR INDEPENDENCE OF EIGENVECTORS CORRESPONDING TO
DISTINCT EIGENVALUES
Exercise 1.
T (x) = λx
(1)
aT (x) = (aλ)x
T1 (x) = λ1 x
(2) (aT1 + bT2 )(x) = aλ1 x + bλ2 x = (aλ1 + bλ2 )x
T2 (x) = λ2 x
Exercise 2.
T (x) = λx
T 2 (x) = T (T (x)) = T (λx) = λT (x) = λ2 x
T n (x) = λn x
T n+1 (x) = T (T n (x)) = T (λn x) = λn λx = λn+1 x
PN
Let P (x) = a j xj
j=0
PN
P (T )(x) = j=0 aj T j (x).
PN PN
If x is an eigenvector, P (T )(x) = j=0 aj T j (x) = j=0 aj λj (x) = P (λ)x.
Exercise 3. V = V2 (R), plane as a real linear space.
π
T = rotation of V through an angle of 2 radians.
T (e1 ) = e2 T 2 (x) = T 2 (x1 e1 + x2 e2 ) = T (x1 e2 + x2 (−e1 )) =
0 −1
T = or =⇒
1 0 T (e2 ) = −e1 = −x1 e1 + x2 (−e2 ) = −x
Exercise 4.
T 2 xλ = λ2 xλ
T 2 − λ2 I = (T + λI)(T − λI)
det(T 2 − λ2 I) = 0 = det(T + λI)det(T − λI) = 0
det(T + λI) or det(T − λI) is zero, so λ or −λ is an eigenvalue of T .
Exercise 5. Let V be the linear space of all real functions differentiable on (0, 1).
Let f ∈ V .
Define g = T (f ) s.t. g(t) = tf 0 (t) ∀ t ∈ (0, 1)
Suppose f is an eigenfunction of T .
g(t) = T (f )(t) = tf 0 (t) = λf (t) =⇒ f (t) = c0 tλ
In solving this ordinary differential equation, λ ∈ R
Exercise 6. V = p(x) of degree ≤ n.
43
p ∈ V , q = T (p) s.t. q(t) = p(t + 1) ∀ t
N
X
p(t) = aj tj
j=0
N
X N
X
T (p(t)) = q(t) = p(x + 1) = aj (t + 1)j = λ aj tj
j=0 j=0
N =0 a0 = λa0 =⇒ λ = 1
a1 (t + 1) + a0 = λ(a1 t + a0 )
t(a1 (1 − λ)) + a1 + a0 (1 − λ) = 0
N =1
if a1 = 0, then a0 = 0 or λ = 1
if λ = 1, a1 = 0
a2 (t + 1)2 + a1 (t + 1) + a0 = λ(a2 t2 + a1 t + a0 )
a2 (t2 + 2t + 1) + a1 (t + 1) + a0 = λ(a2 t2 + a1 t + a0 )
N =2 t2 (a2 (1 − λ)) + t(2a2 + a1 (1 − λ)) + a2 + a1 + (1 − λ)a0 = 0
if a2 = 0, we’re left with N = 1 case
if λ = 1, a2 = 0; we’re left with N = 1 case.
PN PN
Assume j=0 aj (t + 1)j = λ j=0 aj tj , aj = 0 ∀ j = 1, . . . N
N
X +1 N
X +1
aj (t + 1)j = λ aj tj
j=0 j=0
N +1 j N +1
X X j k X
aj t =λ aj tj
j=0
k j=0
k=0
N +1 j
X X j
aj tk − λaj tj = 0
j=0
k
k=0
aN +1 (1 − λ) = 0.
If λ = 1, tN : aN +1 (N + 1) + aN − λaN = 0; aN +1 = 0, and then we could rewrite the equation, and coefficients, as
N th case, which we’ve shown to yield only a0 to be nonzero. Rx
Exercise 7. Let V = linear space of functions continuous on (−∞, ∞) s.t. ∃ −∞ f (t)dt ∀ x ∈ R
Rx
If f ∈ V , let g = T (f ) s.t. g(x) = −∞ f (t)dt
Z x
T (f )(x) = g(x) = f (t)dt = λf (x)
−∞
=⇒ f (x) = λf 0 (x) =⇒ f (x) = c0 eλx
Z x λt x
c0 e c0 eλx c0 eλt
f (t)dt = = − lim
−∞ λ
−∞ λ t→−∞ λ
A limit only exists if λ > 0.
Exercise 8.
Z x
g(x) = T (f )(x) = tf (t)dt = λf (x)
−∞
xf (x) = λf 0 (x)
1 2
f (x) x x2
if λ 6= 0, ln = 2 =⇒ f (x) = c0 e 2λ
f (0) λ
Z x
2 2 2
tc0 et /2λ dt = c0 ex /2λ − lim c0 et /2λ
−∞ t∈−∞
Limit only exists if λ < 0.
44
Exercise 9.
T (f ) = f 00 = λf
f (t) = cn sin nt f (0) = f (π) = 0
f 00 (t) = −n2 cn sin nt = λf =⇒ λn = −n2
Exercise 10. T (x) = (yn ). yn = a − xn ; n≥1
T ((xn )) = (a − xn ) = λ(xn )
The sequences are equal, so a − xn = λxn or a = (λ + 1)xn .
(xn ) is a convergent sequence, so
xn − a = xn − (λ + 1)xn = −λxn must go to zero
So limn→∞ xn = 0, a = 0. Then by a = (λ + 1)xn , λ = −1 for nonzero xn .
λ = −1, (xn ) s.t. lim xn = 0 and xn nonconstant
n→∞
If xn is constant, a = (λ + 1)a =⇒ λ = 0.
λ = 0, xn = a, (xn ) is a constant sequence
Exercise 11. T (x) = λx; T (y) = µy.
T (ax + by) = β(ax + by) = aλx + bµy
a(β − λ)x + b(β − µ)y = 0
Use Thm., Thm. 4.2: Let u1 , u2 , . . . , uk be eigenvectors of linear linear transformation T : S → V .
Assume corresponding eigenvalues λ1 , λ2 , . . . , λk distinct.
Then u1 , u2 , . . . , uk are independent.
=⇒ x, y independent β−λ=β−µ=0
If λ = β, µ 6= β or µ = β, λ 6= β. (given that λ, µ distinct).
T (x) = λx
Exercise 12. Suppose x, y ∈ S, so
T (y) = µy
Suppose λ 6= µ.
Suppose ax + by ∈ S for some a, b ∈ R. Then by definition of S, ax + by is an eigenvector of T .
Then by Exercise 11, a or b is zero. Suppose b = 0.
T (ax + by) = β(ax + by) = aT (x) + bT (y) = aλx + bµy
T (ax) = β(ax) = aλx = λ(ax)
ax nonzero, so β = λ.
So if x ∈ S, so is ax ∈ S, a 6= 0 and T (x) = cx = T (ax) = c(ax)
This must be true ∀ x ∈ S =⇒ T (x) = cx ∀ x ∈ S.
4.8 E XERCISES - T HE FINITE - DIMENSIONAL CASE . C HARACTERISTIC POLYNOMIALS . C ALCULATION OF
EIGENVALUES AND EIGENVECTORS IN THE FINITE - DIMENSIONAL CASE . T RACE OF A MATRIX
Exercise 1.
1 0 1 0
(1) λ1,2 = 1 ζλ=1 = , ; E(1) = 2
0 1 0 1
1 1
(2)
0 1
λ − 1 −1
= (λ − 1)2 = 0
0 λ − 1
1 1 x1 x 1
= 1 =⇒ ζλ=1 = ; E(1) = 1
0 1 x2 x2 0
45
1 0
(3)
1 1
λ − 1 0
−1 = (λ − 1)2 = 0
λ − 1
0
λ = 1, ζλ=1 = ; E(1) = 1
1
1 1
(4)
1 1
λ − 1 −1
−1 = (λ − 1)2 − 1 = (λ − 2)λ = 0; λ = 0, 2
λ − 1
1 1
λ = 2, ζλ=2 = √ ; E(2) = 1
2 1
1 1
λ = 0, ζλ=0 = √ ; E(0) = 1
2 −1
1 a
Exercise 2. ; a > 0, b > 0
b 1
√
p
λ − 1 −a
= (λ − 1)2 − ab = λ2 − 2λ + 1 − ab = 0 2± 4 − 4(1)(1 − ab)
−b λ± = = 1 ± ab
λ − 1 2
√
√ x1 + ax2 = (1 ± ab)x1
1 a x1 x1
= (1 ± ab) =⇒ √
b 1 x2 x2 bx1 + x2 = (1 ± ab)x2
√
√ √
λ± = 1 ± ab; ζλ=1±√ab = √a E(1 ± ab) = 1
± b
cθ −sθ
Exercise 3.
sθ cθ
λ − cθ sθ
−sθ = λ2 − 2λcθ + 1 = 0 =⇒ λ = cθ ± isθ
λ − cθ
1 0
if θ = 2πn, λ = 1, ξλ=1 = ,
0 1
√
cθ −sθ x x 1
if θ 6= 2πn, λ = e−±iθ = e±iθ =⇒ ξλ=e±iθ = 1/ 2
sθ cθ y y ±i
a b
Exercise 4. A =
c d
0 1
P1 =
1 0
|λI − P1 | = λ2 − 1 = 0
0 −i
P2 = |λI − P2 | = λ2 − 1 = 0
i 0
|λI − P3 | = (λ − 1)(λ + 1) = 0
1 0
P3 =
0 −1
λ − a −b
−c = λ2 − (a + d)λ + (ad − bc) = 0 = (λ − λ1 )(λ − λ2 ) = λ2 − (λ1 + λ2 )λ + λ1 λ2
λ − a
∆ = ad − bc = −1 =⇒ a = −d
a b
A= where a2 + bc = 1
c −a
Exercise 5.
46
p
2 (a + d)2 − 4(1)(ad − bc)
(a + d) ±
det(A − λI) = λ − (a + d)λ + (ad − bc) = 0 =⇒ λ =
p 2
if (a + d)2 − 4(ad − bc) > 0, λ real and distinct
p
if (a + d)2 − 4(ad − bc) = 0, λ real and equal
p
if (a + d)2 − 4(ad − bc) < 0, λ complex conjugates
Exercise 6.
1 1 1 1 1 a+b+c=3
a b c 1 = 3 1
d+e+f =3
d e f 1 1
1 1 1 1 1 a−c=0 a=c
a b c 0 = 0 0
d−f =0 d=f
d e f −1 −1
1 1 1 1 1 a=b
a b c −1 = 0 −1
d=e
d e f 0 0
a=b=c=d=e=f =1
Exercise 7.
1 0 0
(1) −3 1 0
4 −7 1
1 − λ 1 − λ
−3 1−λ = 1−λ =0 =⇒ λ = 1
4 −7 1 − λ 1 − λ
1 0
0 x1 x1 x1 = 0
0
−3 1 0 x2 = x2 =⇒ x2 = 0 =⇒ ζλ=1 = 0
4 −7 1 x3 x3 x3 = 1 1
2 1 3
(2) 1 2 3
3 3 20
2 − λ 1 3 1 − λ 1 3 1 − λ 1 3 1 − λ
1
2−λ 3 = −1 + λ 2−λ 3 = 0 3−λ 6 = 3−λ 6
3 3 20 − λ 0 3 20 − λ 0 3 20 − λ 3 20 − λ
(1 − λ)((3 − λ)(20 − λ) − 18) = (1 − λ)(60 − 23λ + λ2 − 18) = (1 − λ)(42 − 23λ + λ2 ) = (1 − λ)(λ − 21)(λ − 2) = 0
=⇒ λ = 1, 2, 21
5 −6 −6
(3) −1 4 2
3 −6 −4
λ − 5 6 6 λ − 2 0 −λ + 2
−1 = (λ − 2)2 (λ − 1)
1
λ−4 −2 = 0 λ−4
−3 6 λ+4 0 6 λ+1
3 4 0
1 1 1
=⇒ ξλ=1 = √ −1 , ξλ=2 = √ 1 , √ 1
19 3 18 1 2 −1
Exercise 8.
47
1
1
(1)
1
1
−λ
1 −λ
1 − λ2
−λ 1 −λ 1 − λ2 2 2
= 1 = (1 − λ ) =⇒ λ = ±1
1
−λ 0
1 −λ 1 0
1
1
(2) =⇒ (λ − 1)2 (λ + 1)2 = 0 so λ = ±1
−1
−1
1
1
(3)
1
1
λ
−1
−1 λ = (λ2 − 1)(λ2 − 1) =⇒ λ = ±1
λ −1
−1 λ
−i
i
(4)
−i
i
−λ −i
2
− −λ −λ −i
= = (λ2 − 1)2
−λ −i i −λ
i −λ
1
−1
=⇒ ((λ + 1)(λ − 1))2 = 0
(5)
1
−1
Exercise 10.
Exercise 11. Let (AB)x = λx
A−1 (AB)x = λ(A−1 x) = BX
Let x = Ay
λ(A−1 Ay) = BAy = λy
So if λ is eigenvalue of AB, λ is also an eigenvalue of BA (A is invertible).
Exercise 13.
Exercise 14.
(1)
N
X N
X N
X N
X
tr(A + B) = (A + B)ii = (aii + bii ) = aii + bii = trA + trB
i=1 i=1 i=1 i=1
(2)
N
X N
X N
X
tr(cA) = (cA)ii = caii = c aii = ctrA
i=1 i=1 i=1
PN PN PN PN PN PN
(3) tr(AB) = j=1 (AB)jj = j=1 k=1 ajk bkj = k=1 j=1 bkj ajk = j=1 (BA)jj = tr(BA)
PN PN
(4) trAT = j=1 (AT )jj = j=1 ajj = trA
48
4.10 E XERCISES - M ATRICES REPRESENTING THE SAME LINEAR TRANSFORMATION . S IMILAR MATRICES .
Exercise 1. Given
1 1 1
A= = xλ=1
0 1 f (λ) = λ2 − 2λ + 1 0
g(λ) = (λ − 1)2
1 0 1 0
B= ζλ=1 = ,
0 1 0 1
Suppose C −1 BC = A.
C −1 C = I = A. But A 6= I.
Contradiction. So @ C invertible s.t. C −1 BC = A
Exercise 2.
1 0
(1) A =
1 3
2 1
ξλ=1 =√
5 −1/2
λ − 1
= (λ − 1)(λ − 3) =⇒
−1 λ − 3
0
ξλ=3 =
1
√
2/ √5 0
C=
−1/ 5 1
1 1√ 0√ 1
indeed, √ C=I
2/ 5 1/ 5 2/ 5 1 3
(2)
1 2
A=
5 4
λ − 1 −2
−5 = λ2 − 5λ + 4 − 10 = (λ − 6)(λ + 1)
λ − 4
1 1
ξλ=−1 = √
−1
1 2 x1 x 2
= −1, 6 1 =⇒
5 4 x2 x2
1 2
ξλ=6 = √
29 5
√ √
1/ √2 2/√29
C=
−1/ 2 5/ 29
√ √
√ √
5/ √29 −2/√ 29 1 2 −1
indeed, 1/(7/ 2 29) C=
1/ 2 1/ 2 5 4 6
2 1
(3) A =
−1 4
λ − 2 −1
= (λ − 3)2
1 λ − 4
Suppose nonsingular C exists, s.t. C −1 AC = 3I =⇒ A = 3I. Contradiction.
(4)
2 1 λ − 2 −1
A= =⇒ = λ2 − 2λ + 1 = (λ − 1)2
−1 0 1 λ
Suppose nonsingular C exists s.t. C −1 AC = I =⇒ A = I. Contradiction.
Exercise 3.
[y1 , y2 ] = [x1 , x2 ]A
B = AC
[z1 , z2 ] = [x1 , x2 ]B −1
A B=C
[z1 , z2 ] = [y1 , y2 ]C = [x1 , x2 ]B = [x1 , x2 ]AC
Exercise 4.
49
(1)
0 0 1 λ 0 −1
0 = λ3 − λ2 − λ + 1
A = 0 1 0 =⇒ 0 λ−1
1 0 0 −1 0 λ
λ3 − λ2 − λ + 1 = (λ − 1)2 (λ + 1)
1 0 1
Note that we could still obtain the following independent eigenvectors: 0 , 1 , 0
1 0 −11
(2)
1 −1 −1
A= 1 3 1 =⇒ f (λ) = λ3 − 5λ2 + 8λ − 4 =
−1 −1 1
= (λ2 − 4λ + 4)(λ − 1) = (λ − 2)2 (λ − 1)
1 1 −1
=⇒ xλ=2 = −1 , xλ=2 = 0 , xλ=1 = 1
0 −1 −1
Exercise 5. Generally, we’ll have
C −1 AC = λI + =⇒ A = λI + C C −1
1 1
a b −1 b 0 1 d −b
C = =
c d 1 d 0 detC −c a
bd −d2
1
=
detC d2 −bd
2 −1
(1) So for A = . Then d = 0, b = 1, c = −1 for C.
0 2
2 1
(2) For A = =⇒ λ = 3. So bd = −1, ad − bc = −1.
−1 4
if b = 1, d = −1, −a − c = −1 a+c=1
if b = −1, d = 1, a + c = −1
Exercise 6.
0 −1 λ 1
−1 = (λ − 1)3
1 =⇒ λ
−1 −3 5 1 3 λ − 3
1
1
=⇒ √ −1
3 −1
Suppose C −1 AC = I. A = CC −1 = 1 so diagonalizing matrix cannot exist for this A.
CHECK this result.
Exercise 7.
(1) ∀ matrix A, we can always consider the characteristic polynomial |λI − A| = f (λ). So ∃ n roots, λj ∈ C If λj 6= 0,
∀ j = 1, . . . , n, then det(C −1 AC) = detA = det(Λ) = j=1 λj 6= 0. So A nonsingular.
Q
If A nonsingular, detA 6= 0, so
detA = det(C −1 AC) = detΛ 6= 0 =⇒ λj 6= ∀ j
(2) A nonsingular,
n
Y n
Y
det(AA−1 ) = detAdetA−1 = detC −1 ACdetD−1 AD = detΛA detΛA = λj bk = 1
j=1 k=1
1
λj , bk distinct, so bk = λj
Exercise 8.
(1) A2 = −1 so A−1 = −A, so A nonsingular.
(2) detA2 = (detA)2 = (−1)n . (detA)2 > 0, so (−1)n = 1; n even.
50
(3) Ax = λx
A2 x = −x = λ2 x λ2 = −1
(4) detA = 1 From fundamental theorem of algebra, roots of the characteristic polynomial must come in complex
conjugate pairs. We already showed that the eigenvalues are purely imaginary. So there must be a whole number of
pairs of complex conjugate eigenvalues that multiply together to get −1. CHECK this result.
5.5 E XERCISES - E IGENVALUES AND INNER PRODUCTS , H ERMITIAN AND SKEW-H ERMITIAN TRANSFORMATIONS ,
E IGENVALUES AND EIGENVECTORS OF H ERMITIAN AND SKEW-H ERMITIAN OPERATORS , O RTHOGONALITY OF
EIGENVECTORS CORRESPONDING TO DISTINCT EIGENVALUES
Exercise 1.
if T (x) = λx, (T (x), y) = (λx, y) = λ(x, y) ∀y ∈ E
if (T (x), y) = (λx, y) ∀y ∈ E
=⇒ (T (x) − λx, y) = 0 ∀y ∈ E
=⇒ T (x) = λx
Exercise 2.
(T (x), y) = (cx, y) = c(x, y)
(x, T (y)) = (x, cy) = c̄(x, y)
since V is a real Euclidean space c ∈ R for (T (x), y), (x, T (y)) ∈ R
Exercise 3.
(1) Assume T : V → V is a Hermitian transformation.
Use induction:
(T x, y) = (x, T y)
(T 2 x, y) = (T x, T y) = (x, T 2 y)
(T n+1 x, y) = (T x, T n y) = (x, T n+1 y)
T −1 is Hermitian since
(T −1 x, y) = (T −1 x, T T −1 y) = (T T −1 x, T −1 y) = (x, T −1 y)
Neat trick, no?
(2) (T (x), y) = −(x, T (y)). Now (T 2 (x), y) = −(T (x), T (y)) = (−1)2 (x, T 2 (y)).
Assume the nth case, (T n (x), y) = (−1)n (x, T n (y)), i.e. T n is Hermitian (skew-Hermitian) if n is even (odd).
Then consider that
(T n+1 (x), y) = −(T n (x), T (y)) = −(−1)n (x, T n+1 (y)) = (−1)n+1 (x, T n+1 (y))
(T −1 (x), y) = (T −1 (x), T T −1 y) = −(T T −1 (x), T −1 (y)) = −(x, T −1 (y))
So T −1 is skew-Hermitian.
Exercise 4.
(1)
((aT1 + bT2 )(x), y) = (aT1 (x) + bT2 (x), y) = a(T1 (x), y) + b(T2 (x), y) = (x, (aT1 )(y)) + (x, (bT2 )y) =
= (x, (aT )(y) + (bT2 )(y)) = (x, (aT1 + bT2 )y)
(2)
(T1 T2 (x), y) = (T2 (x), T1 (y)) = (x, T2 T1 (y))
if T1 T2 = T2 T1 ; T1 T2 is Hermitian
Exercise 5. Let V = V3 (R).
3
X
(T (x), y) = (T (x))j yj = x1 y1 + x2 y2 − x3 y3
j=1
= x1 y1 + x2 y2 + x3 (−y3 ) = (x, T (y))
R1 R1
Exercise 6. 0
f (t)dt = F (1) − F (0) = 0, F (1) = F (0), likewise, for g ∈ V , 0 g(t)dt = G(1) − G(0) = 0, G(1) = G(0).
51
The trick is to use integration by parts
Z 1 Z 1 Z t
(T f, g) = (T f )(t)g(t)dt = f (x)dxg(t)dt =
0 0 0
Z 1 Z 1 Z 1
1
= F (t)g(t)dt − F (0) g(t)dt = F (t)G(t)|0 − f (t)G(t)dt = −(f, T g)
0 0 0
Exercise 7.
(1) Z 1 Z 1 Z −1 Z 1
(T f, g) = T f (t)g(t)dt = f (−t)g(t)dt = f (t)g(−t)(−dt) = f (t)g(−t)dt =
−1 −1 1 −1
= (f, T g)
(2) Z 1 Z 1
(T f, g) = f (t)f (−t)g(t)dt but (f, T g) = f (t)g(t)g(−t)dt
−1 −1
=⇒ Neither symmetric nor skew-symmetric (choose different coefficients for f and g )
(3) Z 1 Z 1 Z 1 Z −1
(T f, g) = T f (t)g(t)dt = (f (t) + f (−t))g(t)dt = fg + − f (t)g(−t)dt =
−1 −1 −1 1
Z 1
= f (t)(g(t) + g(−t))dt = (f, T g)
−1
Hermitian.
(4) Z 1 Z 1 Z 1
(T f, g) = (f (t) − f (−t))g(t)dt = fg − f (−t)g(t)dt =
−1 −1 −1
Z 1 Z −1 Z 1
= fg − f (t)g(−t)(−dt) = f (g(t) − g(−t))dt = (f, T g)
−1 1 −1
Hermitian.
Rb (pf 0 )0 +qf
Exercise 8. Given (f, g) = a f (t)g(t)w(t)dt, T (f ) = w
Z b Z b Z b
(pf 0 )0 (t) + q(t)f (t)
(T f, g) = (T f )(t)g(t)w(t)dt = g(t)w(t)dt = ((pf 0 )0 + qf )g
a a w(t) a
Z b Z b Z b
b
(pf 0 )0 g = (pf 0 )g|a − (pf 0 )g 0 = − (pf 0 )g 0
a a a
p(a)f (a) = 0
since f, g satisfy
p(b)f (b) = 0
Z b Z b Z b
b
(pf 0 )g 0 = pg 0 f |a − (pg 0 )0 f = 0 − (pg 0 )0 f
a a a
b b
((pg 0 )0 + qg)
Z Z
=⇒ (T f, g) = (pg 0 )0 f + qf g = wf = (f, T g)
a a w
Exercise 9. Let V be a subspace of a complex Euclidean space E.
Let T : V → E be a linear transformation and define a scalar-valued function Q on V as follows:
Q(x) = (T (x), x) ∀x ∈ V
(1) T Hermitian.
(T x, x) = (x, T x) = (T x, x) = Q(x) = Q(x) =⇒ Q(x) ∈ R
(2)
(T x, x) = −(x, T x) = −(T x, x) = −Q(x) = Q(x) =⇒ Q(x) pure imaginary
(3)
Q(tx) = (T (tx), tx) = (tT x, tx) (since T is linear)
Q(tx) = t(T x, tx) = t(tx, T x) = tt(T x, x) = ttQ(x)
52
(4)
Q(x + y) = (T (x + y), x + y) = (T x + T y, x + y) = (T x, x + y) + (T y, x + y) =
= (x + y, T x) + (x + y, T y) = (x, T x) + (y, T x) + (x, T y) + (y, T y) =
= (T x, x) + (T x, y) + (T y, x) + (T y, y) = Q(x) + Q(y) + (T (x), y) + (T (y), x)
Q(x + ty) = Q(x) + Q(ty) + (T (x), ty) + (T (ty), x) = Q(x) + ttQ(y) + t(T (x), y) + t(T (y), x)
(5) Suppose T (x) = y 6= 0 for some x ∈ V , y ∈ E, x 6= 0
(T (x), x) = (y, x) = 0
y 6= x, otherwise (x, x) = 0; x = 0. Contradiction. Done.
Q(ax + by) = (T (ax + by), ax + by) = |a|2 (T (x), x) + ba(T (x), y) + ba(T (y), x) + |b|2 (T (y), y)
= ab(T (x), y) + ba(T (y), x) = 0
Let a = 1, b = −1
− (T (x), y) = (T (y), x) = (y, y) > 0
Let ab = i
(T (x), y) = (T (y), x) = (y, y) > 0
Contradiction. Thus y = 0.
(6)
Q(x + ty) = Q(x) + ttQ(y) + t(T (x), y) + t(T (y), x)
Q ∈ R =⇒ Q(x + ty) = Q(x + ty) =⇒ t(y, T (x)) + t(x, T (y)) = t(T (x), y) + t(T (y), x)
=⇒ t((y, T (x)) − (T (y), x)) + t((x, T (y)) − (T (x), y)) = 0
Suppose t = a + bi, a, b arbitrary
(y, T (x)) − (T (y), x) + ((x, T (y)) − (T (x), y)) = 0
=⇒ =⇒ (y, T (x)) − (T (y), x) = 0 so T is Hermitian.
(y, T (x)) − (T (y), x) − ((x, T (y)) − (T (x), y)) = 0
Exercise 10. Legendre polynomials:
1 (n)
Pn (t) = f (t) where fn (t) = (t2 − 1)n
2 n! n
n
(1)
(t2 − 1)fn0 (t) = (t2 − 1)(n)(t2 − 1)n−1 (2t) = 2nt(t2 − 1)n = 2ntfn (t)
(2) Leibniz’s formula. If h(x) = f (x)g(x), prove that the nth derivative of n is given by the formula.
n
X n (k)
h(n) (x) = f (x)g (n−k) (x)
k
k=0
so then
n+1
X
((t2 − 1)fn0 (t))(n+1) =
n+1 2
(t − 1)(k) (fn0 (t))(n+1−k) = (t2 − 1)0 = 2t
k
k=0 (t2 − 1)00 = 2
(n + 1)n (n) (t2 − 1)00 = 0
= (t2 − 1)fn(n+2) (t) + (n + 1)2tfn(n+1) (t) + 2fn (t)
2
(2ntfn (t))(n+1) = (2n)(t(fn (t))(n+1) + (n + 1)(fn (t))(n) )
=⇒ (t2 − 1)fn(n+2) (t) + 2t(n + 1)fn(n+1) (t) + (n + 1)nfn(n) (t) = (2n)(tfn(n+1) (t) + (n + 1)(fn (t))(n) )
1 (n)
(3) Pn (t) = 2n n! fn (t)
=⇒ (t2 − 1)Pn00 + 2t(n + 1)Pn0 + (n + 1)nPn = (2n)(tPn0 + (n + 1)Pn )
=⇒ (t2 − 1)Pn00 + 2tPn0 − (n + 1)nPn = 0 or ((t2 − 1)Pn0 )0 = n(n + 1)Pn
53
5.11 E XERCISES - E XISTENCE OF AN ORTHONORMAL SET OF EIGENVECTORS FOR H ERMITIAN AND
SKEW-H ERMITIAN OPERATORS ACTING ON FINITE - DIMENSIONAL SPACES ; M ATRIX REPRESENTATIONS FOR
H ERMITIAN AND SKEW-H ERMITIAN OPERATORS ; H ERMITIAN AND SKEW-H ERMITIAN MATRICES . T HE ADJOINT OF A
MATRIX ; D IAGONALIZATION OF A H ERMITIAN OR SKEW-H ERMITIAN MATRIX ; U NITARY MATRICES . O RTHOGONAL
MATRICES
Exercise 1.
0 1 2
(1) 1 0 3. Symmetric. Hermitian.
2 3 4
0 i 2
(2) i 0 3 . Skew-Hermitian.
−2 −3 4i
0 i 2
(3) −i 0 3. Skew-symmetric.
−2 −3 0
0 1 2
(4) −1 0 3. Skew-Hermitian. Skew-symmetric.
−2 −3 0
Exercise 2.
(1)
cos θ − sin θ cos θ sin θ 1 0
=
sin θ cos θ − sin θ cos θ 0 1
(2)
cos θ − sin θ r cos α r cos α cos θ − r sin α sin θ cos (α + θ)
= =r
sin θ cos θ r sin α r cos α sin θ + r sin α cos θ sin (α + θ)
Exercise 3.
1 0 0
(1) 0 1 0 (reflection in the xy-plane).
0 0 −1
1 1 1
1 0 = 0
−1 0 0
1 1 1 1 0 0
1 1 = 1 1 1 = 1
−1 −1 1 −1 0 0
1 0 0
1 0 = 0
−1 1 −1
1 0 0
(2) 0 −1 0 (reflection through the x-plane).
0 0 −1
1 1 1
−1 0 = 0
−1 0 0
1 1 1 1 0 0
−1 −1 = 1 −1 1 = − 1
−1 −1 1 −1 0 0
1 0 0
−1 0 = 0
−1 1 −1
54
−1 0 0
(3) 0 −1 0 (reflection through the origin).
0 0 −1
−1 −1 1 −1
−1 −1 = 1 −1 i, j, k = −i, j, k)
−1 −1 1 −1
1 0 0
(4) 0 cos θ − sin θ (rotation about the x-axis).
0 sin θ cos θ
1 0 0 1 0 0 1 0 0
0 cos θ − sin θ 0 cos θ + sin θ = 0 1 0
0 sin θ cos θ 0 − sin θ cos θ 0 0 1
1 0 0 1 1
0 cos θ − sin θ 0 = 0
0 sin θ cos θ 0 0
1 0 0 0 0
0 cos θ − sin θ 1 = cos θ
0 sin θ cos θ 0 sin θ
1 0 0 0 0
0 cos θ − sin θ 0 = − sin θ
0 sin θ cos θ 1 cos θ
−1 0 0
(5) 0 cos θ − sin θ (rotation about x-axis followed by reflection in the yz-plane).
0 sin θ cos θ
−1 0 0 −1 0 0 1 −1 1 −1 0 0
0 cos θ − sin θ 0 cos θ sin θ = 1 1 cθ −sθ = 0 cos θ − sin θ =
0 sin θ cos θ 0 − sin θ cos θ 1 1 sθ cθ 0 sin θ cos θ
= (reflection in the yz-plane)(rotation about x-axis)
Exercise 4. A real orthogonal matrix A is called proper if detA = 1, and improper if detA = −1.
(1)
2 ac + bd = 0
a + b2 ac + bd
a b a c
= =⇒ a2 + b2 = 1
c d b d ac + bd c2 + d2
ad − bc = 1
a = cos θ
b = sin θ
−c
=⇒ cos2 θ − sin θc = 1 =⇒
sin θ c = − sin θ
d = cos θ
(2)
1 1 1 1 0
= det = −1
−1 −1 1 0 −1
−1 −1 1 −1 0
= det = −1
1 1 1 0 1
From previous part, all improper 2 × 2 matrices:
cos θ sin θ
sin θ − cos θ
Exercise 13. If A is a real skew-symmetric matrix, prove that both I − A and I + A are nonsingular and that (I − A)(I + A)−1
is orthogonal.
Note: Notice the difference between skew-Hermitian and skew symmetric and use eignevalue eqn. and one-to-one.
55
Skew-Hermitian matrices must be square matrices (from how Skew-Hermitian operators, T , are defined as T : V → V ).
For skew-symmetric matrices, eigenvalues must equal zero, since −λ = λ
C −1 AC = Λ = 0
det(1 ± A) = det(C −1 C)det(1 ± A) = det(C −1 )det(1 ± A)detC =
= det(1 ± C −1 AC) = det(1 ± 0) = 1
To prove orthogonality of (I − A)(I + A)−1 , use (AB)T = B T AT extensively.
We know that A is real skew-symmetric, so that A = −AT .
(1 + A)(1 + A)−1 = 1
((1 + A)(1 + A)−1 )T = ((1 + A)−1 )T (1 + AT ) = 1T = 1
Thus,
(1 − A)(1 + A)−1 ((1 + A)−1 (1 − A))T =
= (1 − A)(1 + A)−1 (1 − AT )((1 + A)−1 )T = (1 − A) (1 + A)−1 (1 + A) ((1 + A)−1 )T =
= (1 − A)((1 + A)−1 )T = (1 + AT )((1 + A)−1 )T = (1 + A)T ((1 + A)−1 )T = ((1 + A)−1 (1 + A))T = 1T = 1
Note that we have ((1 − A)(1 + A)−1 )T = ((1 + A)−1 (1 − A))T because if (1 + A)−1 = B,
(1 + A)B = B + AB = 1
=⇒ BA = AB
B(1 + A) = B + BA = 1 (since a left inverse is a right inverse)
(1 − A)(1 + A)−1 = (1 − A)B = B − AB = B − BA = B(1 − A) = (1 + A)−1 (1 − A)
Exercise 14.
(1) Counterexample:
1 + ei2π/3
i2π/3
1 e
A= B= (A + B) =
e−i2π/3 1 1 + e−i2π/3
−i2π/3
1 + e−i2π/3
1 e
A∗ = B∗ = A∗ + B ∗ = (A + B)∗ =
ei2π/3 1 1 + ei2π/3
2 + ei2π/3 + e−i2π/3
(A + B)(A∗ + B ∗ ) =
2 + ei2π/3 + e−i2π/3
(2) If A and B are unitary, then AB is unitary.
AA∗ = BB ∗ = 1
(AB)(AB)∗ = (AB)B ∗ A∗ = A(BB ∗ )A∗ = A1A∗ = 1
(3)
(4)
5.15 E XERCISES - Q UADRATIC FORMS , R EDUCTION OF A REAL QUADRATIC FORM TO A DIAGONAL FORM ,
A PPLICATIONS TO ANALYTIC GEOMETRY
4 2
Exercise 1. 4x21 + 4x1 x2 + x22 . =A
2 1
4 − λ 2
= 4 − 5λ + λ2 − 4 = λ(λ − 5)
2 1 − λ
√ √
2/√5 1/ √5
ξλ=5 = ; ξλ=0 =
1/ 5 −2/ 5
1 2 1
C=√
5 1 −2
0 1/2
Exercise 2. x1 x2 A= . λ = 21 , − 12 .
1/2 0
√ √
1/√2 1/ √2
ξλ=1/2 = , ξλ=−1/2 =
1/ 2 −1/ 2
56
1 1
Exercise 3. x21 + 2x1 x2 − x22 . A=
1 −1
34 −12
Exercise 4. 34x21 − 24x1 x2 + 41x22 . A= .
−12 41
λ − 34 12
= λ2 − 75λ + 34(41) − 144 = λ2 − 75λ + 1250. λ = 50, 25
12 λ − 41
1 3
ξλ=50 =
5 −4
1 3 4
C=
5 −4 3
1 4
ξλ=25 =
5 3
Exercise 5. x21 + x1 x2 + x1 x3 + x2 x3 .
1 1 1 λ − 1 −1 −1 λ −1 0
0 = (λ + 1)(λ2 − 2λ − 1)
A = 1 0 1 =⇒ −1 λ −1 = −1 λ − 2
1 1 0 −1 −1 λ 0 λ + 1
0
1
ξλ=−1 =√ 1 √ √
2 −1 0√ 2/2 − 2/2
√ C = 1/ √2 1/2 1/2
1 ± 2 −1/ 2 1/2 1/2
ξλ=1±√2 = 1
2
1
Exercise 6. 2x21 + 4x1 x3 + x22 − x23
2 2 λ − 2 −2
A= 1 =⇒ λ−1 = (λ − 1)(λ − 3)(λ + 2)
2 −1 −2 λ + 1
ξλ=1 = (0, 1, 0)
√ √
1 0 2/ 5 1/ 5
ξλ=3 = √ (2, 0, 1)
5 =⇒ C = 1 0√ 0√
1 0 1/ 5 −2/ 5
ξλ=−2 = √ (1, 0, −2)
5
Exercise 7. 3x21 + 4x1 x@ + 8x1 x3 + 4x2 x3 + 3x23 .
3 2 4
A = 2 0 2
4 2 3
λ − 3 −2 −4 λ + 1 −2 0 λ+1 −2 0
det(λI − A) = −2 λ −2 = −2λ − 2 λ −2λ − 2 = −2λ − 2 λ−4 0 =
−4 −2 λ − 3 0 −2 λ+1 0 −2 λ + 1
= (λ + 1)(λ − 4)(λ + 1) + 2(λ + 1)(−2λ − 2) = (λ + 1)(λ2 − 3λ − 4 − 4λ − 4) = (λ + 1)2 (λ − 8)
57
3 24 x1 x1 4x1 + 2x2 + 4x3 = 0
2 02 x2 = −1 x2 =⇒
2x1 + x2 + 2x3 = 0
4 23 x3 x3
1
1
=⇒ ξλ=−1 = √ 0
2 −1
3 2 4 x1 x1 −5 2 4 0 −18 9 2 −1
2 0 2 x2 = 9 x2 =⇒ 2 −8 2 = 1 −4 1 = 1 0 −1
4 2 3 x3 x3 4 2 −5 0 18 −9
2
1
=⇒ ξλ=8 = 1
3
2
e1 e2 e3 1
1
1 0 −1 = (1, −4, 1) =⇒ ξλ=−1 = √ −4
2 1 2
3 2 1
√ √
1/ 2 1/3 √2 2/3
C = 0√ −4/3√ 2 1/3
−1/ 2 1/3 2 2/3
Exercise 8. y 2 − 2xy + 2x2 − 5 = 0.
2 −1 λ − 2 1
=⇒ = λ2 − 3λ + 1 = 0
−1 1 1 λ − 1
√
3± 5
=⇒ λ =
2
1 1√
√ ξλ= 3+√5 = q √ 1− 5
2 5− 5
2 −1 x1 3 ± 5 x1 2
2
= =⇒
−1 1 x2 2 x2
1 1√
ξλ= 3−√3 = p √ 1+ 5
2
5 + 52 2
q 1√ 1
q √
5− 5 5+ 5
=⇒ C = q 2 √ q 2
√
2√ 1− 5 2√ 1+ 5
5− 5 2 5+ 5 2
√ √
3+ 5 2 3− 5 2
=⇒ x + y =5
2 2
Ellipse centered about (0, 0).
Exercise 9. y 2 − 2xy + 5x = 0
0 −1
=A
−1 1
p √
λ 1 1± 1 − 4(1)(−1) 1± 5
|λI − A| =
= λ2 − λ − 1 =⇒ λ = =
1 λ − 1 2 2
√ " #
0 −1 x 1+ 5 x 1 1√
= =⇒ ξλ= 1+√5 =q √ 1+ 5
−1 1 y 2 y 2 5+ 5 −2
2
√
0 −1 x 1− 5 x 1 1√
= =⇒ ξ √
λ= 1−2 5
=q √ −1+ 5
−1 1 y 2 y 5− 5 2
2
q q
2√ 2√
5+
5√ 5−
5 √
=⇒ C = q
2√ 1+ 5
q
2√ −1+ 5
5+ 5 −2 5− 5 2
58
y 2 − 2xy + 5x = 0 =⇒
√ √ s s ! √ s √ s
1+ 5 2 1− 5 2 2 2 1+ 5 2 2 1− 5 2 2
x + y +5 √ x+ √ y = x +5 √ x+ y +5 √ y=0=
2 2 5+ 5 5− 5 2 5+ 5 2 5− 5
√ s ! √ s !
1+ 5 2 2 1 − 5 2 2
= x2 + 5 √ √ x + y2 + 5 √ √ y =0
2 5+ 5 1+ 5 2 5− 5 1− 5
√ s !2 √ ! s !2
1+ 5 2 1 1− 5 2 1
=⇒ x+5 √ √ + y+5 √ √ =
2 5+ 5 1+ 5 2 5− 5 1− 5
√ √
1 + 5 2 (1 − 5)52
= √ 5 + √
40 + 16 5 40 − 16 5
q
−5 5+2√5 1+1√5 5
CY = X =⇒ C q = 25
−5 5−2√5 1−1√5 2
Ellipse centered at (5/2, 5/2).
Exercise 10. y 2 − 2xy + x2 − 5x = 0.
1 −1
A=
−1 1
1 − λ −1
−1 = 1 − 2λ + λ2 − 1 = λ(λ − 2) =⇒ λ = 0, 2
1 − λ
1 −1 x1 1 1
= 0 =⇒ ξλ=0 = √
−1 1 x2 2 1
1 1
similarly, ξλ=2 = √
2 −1
1 1 1 5x2 5
C=√ =⇒ 2x22 − √ = √ x1
2 1 −1 2 2
√ 2
2 2 5 5
=⇒ x2 − √ = x1 + √
5 4 2 8 2
" #
5
− 8√
x1 x 2
C = =C 5
x2 y √
4 2
5 −15
=⇒ (x, y) = ,
16 16
5 −15
The vertex of the parabola in (x, y) coordinates is 16 , 16 .
Exercise 11. 5x2 − 4xy + 2y 2 − 6 = 0.
5 −2 5 − λ −2
=A = 10 − 7λ + λ2 − 4 = (λ − 6)(λ − 1)
−2 2 −2 2 − λ
4 −2 x1 1 1
λ=1 =0 ξλ=1 = √
−2 1 x2 5 2
−1 −2 x1 1 2
λ=6 = 0 ξλ=6 = √
−2 −4 x2 5 −1
x21 + 6x22 = 6
x21
+ x22 = 1
6
Ellipse centered at (0, 0) in both sets of coordinates.
Exercise 12. 19x2 + 4xy + 16y 2 − 212x + 104y = 356.
59
19 2 19 − λ
2
= (λ − 15)(λ − 20)
2 16 216 − λ
1 −1
ξλ=15 = √
5 2
1 2
ξλ=20 = √
5 1
−1
Y C = X so
1 −1 2 −x1 + 2x2 2x1 + x2
[x y] = [x1 x2 ] √ = √ , √
5 2 1 5 5
2 2 −x 1 + 2x 2 2x 1 + x 2
=⇒ 15x1 + 20x2 + −212 √ + 104 √ = 356
5 5
2 2
14 √8
x1 + √ 5
x 2 − 5
=⇒ 403
+ 1209
=1
5 20
Suppose we want to know what the center is in terms of the original (x, y) coordinates. Use C.
x1 1 −1 2 x1 1 −x1 + 2x2
C =√ =√ =
x2 5 2 1 x2 5 2x1 + x2
(x1 ,x2 )= −14 8
√ ,√
5 5 6
−−−−−−−−−−−−→=
−4
Thus, we have an ellipse centered at (6, −4).
Exercise 13. 9x2 + 24xy + 16y 2 − 52x + 14y = 6
9 12 1 3 1 4
ξλ=25 = ; ξλ=0 =
12 16 5 4 5 −3
−1 1 4 −3 1
X =YC =⇒ [x y] = [x1 x2 ] = [4x1 + 3x2 , −3x1 + 4x2 ]
5 3 4 5
2 4x1 + 3x2 −3x1 + 4x2
=⇒ 25x2 − 52 + 14 =6
5 5
1 12 1
=⇒ (x2 − )2 = + x1
2 5 5
To get the center in terms of the (x, y) original coordinates,
x 1 4 3 x1 1 4x1 + 3x2
C 1 = =
x2 5 −3 4 x2 5 −3x1 + 4x2
−1 2
(x1 ,x2 )=( 5 , 5 ) 2/25
−−−−−−−−−−→=
11/25
2 11
Thus we have a parabola centered at 25 , 25 .
Exercise 14. 5x2 + 6xy + 5y 2 − 2 = 0
5 3
A=
3 5
1 1 1 1
λ = 2, 8 ξλ=2 =√ ξλ=8 = √
2 −1 2 1
2 2
=⇒ 2x1 + 8x2 − 2 = 0
=⇒ x21 + 4x22 = 1
Thus we have an ellipse centered about the origin in both coordinate axes.
Exercise 15. x2 + 2xy + y 2 − 2x + 2y + 3 = 0
60
1 1 1 − λ 1
A= = λ(λ − 2)
1 1 1 1 − λ
Directly from the characteristic function,
1 1 x1 1 1
=0 √ = ξλ=0
1 1 x2 2 −1
=⇒
−1 1 x1 1 1
= 0 ξλ=2 = √
1 −1 x2 2 1
(
x = √12 (x1 + x2 )
1 1 −1
Y C T = X = [x y] = [x1 x2 ] √ =
2 1 1 y = √12 (−x1 + x2 )
2 2
=⇒ 2x22 + − √ (x1 + x2 ) + √ (−x1 + x2 ) + 3 = 0
2 2
√ √
2 2 3 2
x + = x1
2 2 4
x1 1 1 1 x1 1 x1 + x2 x
C =√ =√ =
x2 2 −1 1 x2 2 −x 1 + x 2 y
√
3 2 3 −3
For vertex at 4 ,0 in (x1 , x2 ) coordinates, vertex has 4, 4 as (x, y) coordinates.
Exercise 16. 2x + 4xy + 5y 2 − 2x − y − 4 = 0.
2
2 2
A=
2 5
2 − λ 2
= (2 − λ)(5 − λ) − 4 = 10 − 7λ + λ2 − 4 = 6 − 7λ + λ2 = (λ − 6)(λ − 1)
2 5 − λ
1 2 1 1
ξλ=1 = √ ξλ=6 = √
5 −1 5 2
1 2 1
C=√
5 −1 2
1
Y C T = X = [2x1 + x2 , −x1 + 2x2 ] √
5
2x 1 + x 2 −x 1 + 2x 2
x21 + 6x22 − 2 √ − √ −4=0
5 5
2 2
3 1
3x 4x x1 − 2√ 5
x 2 − √
3 5
1 2
=⇒ x21 − √ + 6x22 − √ = 4 =⇒ 55
+ 55
=1
5 5 12 72
To find the center of the ellipse in terms of (x, y),
x1 1 2 1 x1 1 2x1 + x2 x
C =√ =√ =
x2 5 −1 2 x2 5 −x 1 + 2x 2 y
3 1
(x1 ,x2 )= 2√ , √
5 3 5 2 −1
−−−−−−−−−−−−−→ ,
3 6
3 1 2 1
For the center of an ellipse at , √ in (x1 , x2 ) coordinates, the center of the ellipse in (x, y) coordinates is 3, −6 .
√
2 5 3 5
2 2
Exercise 17. x + 4xy − 2y − 12 = 0
1 2 1 − λ 2
= (1 − λ)(−2 − λ) − 4 = (λ + 3)(λ − 2).
2 −2 2 −2 − λ
61
√1
2 √1
−1
λ = 2, −3 ξλ=2 = ξλ=−3 = .
5 1 5 2
1 2 −1
C=√
5 1 2
1
Y C −1 = X = [x y] = [x1 x2 ]C = √ [2x1 − x2 , x1 + 2x2 ]
5
x21 x2
=⇒ 2x21 − 3x22 − 12 = 0 =⇒ − 2 =1
6 4
Exercise 18. xy + y − 2x − 2 = 0
0 1/2 −λ 1/2 1
A= = λ2 − = 0
1/2 0 1/2 −λ 4
1 1 1 1
λ = ±1/2 ξλ=1/2 =√ ξλ=−1/2 = √
2 1 2 −1
1 1 1
C=√
2 1 −1
1 1 1 x1 + x2 x1 − x2
Y C −1 = X = [x y] = [x1 x2 ] √ = √ , √
2 1 −1 2 2
1 −1 2 x1 − x2 x1 + x2
=⇒ x21 + x + √ −2 √ =2
2 2 2 2 2
2 2
1 3 1 9
=⇒ x1 − √ + − x2 + √ =4+ − =0
2 2 2 2
Suppose two lines are the asymptotic limit of a hyperbola. Then these lines are “hyperbolas.”
1 3
± x1 − √ = x2 + √
2 2
of this “hyperbola” is in terms of coordinates in the original x, y axis (we already have them
If we want to get what the center
for (x1 , x2 ) = Y and that is √12 , − √32 , then apply C as a transformation.
1 1 1 1 1 −1
C(x1 , x2 ) = √ √ =
2 1 −1 2 −3 2
The center is (−1, 2) in (x, y) coordinates.
Exercise 19. 2xy − 4x + 7y + c = 0.
0 1
1 0
1 1
ξλ=1 = √
2 1
λ −1 1 1 1
−1 = λ2 − 1 = 0 =⇒ C = √
λ 2 1 −1
1 1
ξλ=−1 = √
2 −1
!
√1 (y1 + y2 )
1 1
CY =X 1 y1 x
−−−−−→ √ 2 = 1 =
√1 (y1 − y2 )
2 1 −1 y2 2
x2
1 1 3 11
x2 + −y 2 − 4 √ (x + y) + 7 √ (x − y) + c = x2 + √ x + −y 2 − √ y + c = 0
2 2 2 2
2 2
3 11
x+ √ − y+ √ + c = 0 =⇒ c = −14
2 2 2 2
Exercise 20. Note that
a b/2
ax2 + bxy + cy 2 = 1 = XAX T whereA =
, X= x y
b/2 c
62
A symmetric, by theorem, A can be diagonalized into its eigenvalues. By thm., XAX T = Y ΛY T where Y is an orthonormal
coordinate transformation.
b2
|λ − A| = (λ − a)(λ − c) − = (λ − λ+ )(λ − λ− )
4
b2
So λ+ λ− = ac − 4 (this had been the smart way to see this: you can do algebra to get)
p
(a + c) ± (a − c)2 + b2
λ± =
2
Then using eigenvectors as basis,
z2 w2
ax2 + bxy + cy 2 = + =1
1/λ+ 1/λ−
From geometry, the area of an ellipse is πab, where a, (b) is the half of semi-major (minor) axis
s s
1 1 1 2π
ellipse area = π = πq =√
λ+ λ− ac − b
2 4ac − b2
4
5.20 E XERCISES - E IGENVALUES OF A SYMMETRIC TRANSFORMATION OBTAINED AS VALUES OF ITS QUADRATIC
FORM ; E XTREMAL PROPERTIES OF EIGENVALUES OF A SYMMETRIC TRANSFORMATION ; T HE FINITE - DIMENSIONAL
CASE ; U NITARY TRANSFORMATIONS
Exercise 1.
(1)
(T (x), T (x)) = (x, x) = |c|2 (x, x)
(x, x) > 0 if x 6= 0, so |c|2 = 1
(2) V one-dim. T : V → V .
T unitary, T linear. x ∈ V ; x = ae1 ; T (x) = T (ae1 ) = aT (e1 ). T (e1 ) ∈ V ; T (e1 ) = µe1 .
T (x) = aµe1 = µx. Since T unitary, by above |µ| = 1. If V real, µ real, so µ = ±1.
Exercise 2.
T
(1) A real, orthogonal. AA∗ = AA = AAT = 1. Thus, A unitary.
=⇒ eigenvalue λ of A s.t. |λ| = 1.
If λ ∈ R, λ = ±1.
(2)
(3) If n = 2s + 1 odd, suppose λ1 is eigenvalue of A. If λ1 real, done. If λ1 non-real,
then λ1 is an eigenvalue (by previous part).
Continue, until nth eigenvalue (we’ve already checked s pairs of eigenvalues to be non-real). If λn non-real, λn is
an eigenvalue. Then there are 2s + 2 eigenvalues. But we’re given that n odd. Contradiction. Thus λn real.
Exercise 3. T or thogonal. Then m(T ) = A has at least one real eigenvalue, |λn | = 1.
s
!
Y
2
Given detA = 1, detA = 1 = |λi | λn = (1)λn
i=1
Since suppose there are s complex eigenvalues. Then there are s complex conjugate Qs eigenvalues.
Then there are at most
2
n − 2s = (odd − even) = odd number of real eigenvalues. Since detA = 1, and i=1 |λi | = 1 already (T orthogonal),
there can only be an even number of real eigenvalues equal to −1. Then there must be at least one eigenvalue equal to 1.
Exercise 4. A real, orthogonal, then A unitary. Then for eigenvalues λ of A, |λ| = 1. Consider all complex λ of A; they come
Qs
in complex conjugate pairs, and so if there are s conjugate pairs, i=1 |λs |2 = 1.
Consider all real eigenvalues of A. Then λ = ±1. If −1 is an eigenvalue of multiplicityQ of k, then there are k diagonal entries
s
of −1 for diagonalized A. Thus, all possible eigenvalues are considered, so detA = i=1 |λs |2 (1)(−1)k = 1(1)(−1)k =
(−1)k
Exercise 5. Given that T linear and norm-preserving,
63
(T (x + by), T (x + by)) = kT (x)k2 + b(T (y), T (x)) + b(T (x), T (y)) + |b|2 kT (t)k2 = kT (x + by)k2 =
= kx + byk2 = kxk2 + b(x, y) + b(y, x) + |b|2 kyk2
kT (x)k2 = kxk2
as well =⇒ b ((T (y), T (x)) − (y, x)) + b ((T (x), T (y)) − (x, y)) = 0
kT (y)k2 = kyk2
b, b are independent since b = s + ti and s, t are two arbitrary real numbers. So (T (x), T (y)) = (x, y), so T unitary.
Exercise 6. T : V → V unitary, Hermitian.
(T (x), y) = (x, T (y)) (Hermitian)
2 2
(T (x), y) = (T (x), T (y)) = (x, T (y)) = (x, y) =⇒ (T 2 (x) − x, y) = 0
Let y = x.
((T 2 − I)(x), x) = Q1 (x) = 0 ∀ x ∈ V
Then T − I = 0 (as previously shown for Q1 (x) = 0 ∀ x ∈ V ), or T 2 − I.
2
Exercise 7. (e1 , . . . , en ), (u1 , . . . , un ) are 2 orthonormal bases for Euclidean space V .
Pn
ej ∈ V so ej k=1 ajk uk .
n n
! n X n
X X X
(ei , ej ) = (ej , ei ) = ail ul , ajk uk = ail ajk (ul , uk ) =
l=1 k=1 l=1 k=1
n
X
= aik ajk
k=1
=⇒ A is unitary, T s.t. m(T ) = A is unitary (isomorphism).
1 1
a 2i 2 a(2i − 1)
Exercise 8. ia 12 (1 + i) 1
2 a(1 − i)
−1 1
a 2 2 a(2 − i)
n
X 3
X
aki akj = aki akj = (ei , ej )
k=1 k=1
1
a2 + ia(−ia) + a2 = 3a2 = 1 =⇒ a2 =
3
(a, ia, a), 12 i, 12 (1 + i), −1
a
2 , 2 (2i − 1, 1 − i, 2 − i) are orthogonal to each other through (x, y) inner product on complex
p
Euclidean V . If a = ± 1/3, columsn of A will be normalized.
=⇒ AT A = I, so A unitary.
Exercise 9. A skew-Hermitian, Λ = C ∗ AC.
n
Y
det(1 ± A) = det(C ∗ C)det(1 ± A) = det(1 ± C ∗ AC) = det(1 ± A) = (1 ± λj )
j=1
If λj ∈ C, λj purely imaginary and λj is also an eigenvalue.
(1 ± λj )(1 ± λj ) = 1 ± (λj + λj ) + |λj |2 = 2
If λj ∈ R, λj = 0. 1 ± λj = 1
=⇒ det(1 ± A) 6= 0 so 1 ± A nonsingular
−1
Let B = (1 + A) . Use the fact that a left inverse is also a right inverse (theorem) extensively.
(1 + A)B = B + AB = 1 ((1 + A)B)∗ = B ∗ (1 + A∗ ) = B ∗ + B ∗ A∗ = 1
=⇒ AB = BA =⇒ B ∗ A∗ = A∗ B ∗
B(1 + A) = B + BA = 1 (B(1 + A))∗ = (1 + A∗ )B ∗ = B ∗ + A∗ B ∗ = 1
so
B(1 − A) = B − BA = B − AB = (1 − A)B
∗
Thus, using A = −A , since A skew-Hermitian,
(1 − A)(1 + A)−1 ((1 − A)(1 + A)−1 )∗ = (1 − A)B((B(1 − A))∗ ) = (1 − A)B(1 − A∗ )B ∗ = (1 − A)B(1 + A)B ∗ =
= (1 − A)B ∗ = (1 + A∗ )B ∗ = 1
(1 − A)(1 + A)−1 unitary.
Exercise 10. A unitary, I + A nonsingular. Let (1 + A)−1 = B. Using this fact
64
B(1 + A) = B(AA∗ + A) = BA(1 + A∗ ) = 1 = 1∗ = (1 + A)(A∗ B ∗ )
=⇒ A∗ B ∗ = B ∗ A∗ = B
Then
((1 − A)B)∗ = B ∗ (1 − A∗ ) = B ∗ − A∗ B ∗ = (A − 1)A∗ B ∗ = −(1 − A)B ∗ A∗ = −(1 − A)B
Thus (1 − A)B is skew-Hermitian.
Exercise 11. A Hermitian, so A = A∗ . Let B = (A − i)−1
B(A − i) = 1 = (A − i)B =⇒ AB = BA
(B(A − i)) = 1 = (A∗ + i)B ∗ = A∗ B ∗ + iB ∗ = B ∗ A∗ + iB ∗
∗
=⇒ A∗ B ∗ = B ∗ A∗
Then
B(A + i)(B(A + i))∗ = (A + i)B(A + i)∗ B ∗ = (A + i)B(A − i)B ∗ = (A∗ + i)1B ∗ = 1
Exercise 12. A unitary, so by theorem, there exists a complete set of orthonormal eigenvectors that form a basis for V ,
{u1 , . . . , un }.
Suppose A was defined in the {e1 , . . . , en } basis. Then they are related through some matrix C (most general assumption to
make):
n X
X n
[u1 , . . . , un ] = [e1 , . . . , en ]C =⇒ uj = cij ei
i=1 i=1
n n
! n X
n n
X X X X
(uj , uk ) = cij ei , clk el = cij clk (ei , el ) = cij cik =
i=1 l=1 i=1 l=1 i=1
n
X
= (C ∗ )kl clj = (uk , uj )
l=1
Hence C ∗ C = 1, so C is unitary.
Recall what the entries of matrix A are, evaluated from the inner product in a certain chosen basis:
n
X
Ael = alm em
m=1
n
!
X
(Aek , el ) = akm em , el = akl
m=1
Thus,
n
X n
X n
X n X
X n n X
X n
(CAC ∗ )ij Cik (AC ∗ )kj = Cik akl C jl = cik akl cjl = cik (Aek , el )cjl =
k=1 k=1 l=1 k=1 l=1 k=1 l=1
n n
!
X X
= A cik ek , cjl el = (Aui )
k=1 l=1
Exercise 13. A square matrix is called normal if AA∗ = A∗ A. Determine which of the following types of matrices are normal.
(1) Hermitian matrices. AA∗ = A∗ A since A = A∗
(2) Skew-Hermitian matrices. AA∗ = −A∗ (−A) = A∗ A
(3) Symmetric matrices.
(4) Skew-symmetric matrices.
(5) Unitary matrices.
(6)
Exercise 14. If A is a normal matrix (AA∗ = A ∗ A) and if U is a unitary matrix, prove that U ∗ AU is normal.
(U ∗ AU )(U ∗ AU )∗ = U ∗ AU U ∗ A∗ U = U ∗ AA∗ U = U ∗ A∗ AU = U ∗ A∗ U U ∗ AU =
= (U ∗ AU )∗ (U ∗ AU )
65
6.21 E XERCISES - L INEAR EQUATIONS OF SECOND ORDER WITH ANALYTIC COEFFICIENTS , T HE L EGENDRE
EQUATION , T HE L EGENDRE POLYNOMIALS , RODRIGUES ’ FORMULA FOR THE L EGENDRE POLYNOMIALS
Exercise 1.
a. From (6.35) (1 − x2 )y 00 − 2xy 0 + α(α + 1)y = 0.
v0
If α = 0, (1 − x2 )y 00 − 2xy 0 = 0, y 0 = v, so v = 2x
1−x2 =⇒ ln v
v0 = − ln (1 − x2 ) or v
v0 = 1
1−x2
Z
1 x 1 v0 1+x
y − y0 = +v0 dx = +v 0 ln √ + √ = ln
1 − x2 1 − x2 1 − x2 2 1−x
x = sin θ
We did this integral by considering the following: . So
dx = cos θdθ
Z Z Z
1 c(θ)dθ
dx = = sec θdθ = ln (tan θ + sec θ)
1 − x2 cos2 θ
since (ln (tan θ + sec θ))0 = tan θ+sec
1 2
θ (sec θ + tan θ sec θ)
Now by Apostol’s notation,
u1 is the power series solution with a0 = 1, a1 = 0
u2 is the power series solution with a0 = 0, a1 = 1
My notation:
u1 is the power series solution with a0 = 0, a1 = 1
u2 is the power series solution with a0 = 1, a1 = 0
Since (1 − x2 )y 00 − 2xy 0 = 0, 1 − x2 , −2x analytic (have power series representation).
∞
X
y= an xn
n=0
X∞
y0 = nan xn−1
n=1
X∞ ∞
X
y 00 = n(n − 1)an xn−2 = (n + 2)(n + 1)an+2 xn
n=2 n=0
∞
X ∞
X
2a2 + 2(3)a3 x + ((n + 2)(n + 1)an+2 − n(n − 1)an )xn = 2 nan xn
n=2 n=1
=⇒ ∞
X
or 2a2 + 2(3a3 − a1 )x = ((n + 2)(n + 1)an+2 − n(n + 1)an )xn = 0
n=2
nan
So a2 = 0, a3 = a1 /3, an+2 = n+2 .
(2m − 1)a2m−1 (2m − 1) 2m − 3 1
a2m+1 = = a2m−3 = a1
2m + 1 2m + 1 2m − 1 2m + 1
∞
X x2m+1
=⇒ y = a1
m=0
2m + 1
Indeed, since
Z
1
Z X X (−1)j xj+1
= ln (1 + x) = (−x)j =
1+x j+1
Z Z X j+1
1 X x
= − ln (1 − x) = (xj ) =
1−x j+1
So that
∞
1 1 X 2x2m+1
(ln (1 + x) − ln (1 − x)) =
2 2 m=0 2m + 1
66
b.
1 1 1 1 1−x+1+x 1
u02 = + = 2
=
1+x 1−x
2 2 1−x 1 − x2
1 −1 1
u002 = +
2 (1 + x)2 (1 − x)2
1 −(1 − 2x + x2 ) + 1 + 2x + x2
2 00 1 −1(1 − x) 1 + x 2x
(1 − x )u2 = + = =
2 1+x 1−x 2 1 − x2 1 − x2
Exercise 2. Let α = 1. Then (1 − x2 )y 00 − 2xy 0 + 2y = 0.
x 1+x
f (x) = 1 − log
2 1−x
1 1 + x x
f 0 (x) = − log −
2 1 − x 1 − x2
−1 1 −2x2 −2 −2x2
f 00 (x) = 2
− 2
+ 2 2
= 2
+
1−x 1−x (1 − x ) 1−x (1 − x2 )2
−2 −2x2 2x2 2x4 2x2 1+x 2x2 1+x
=⇒ + + + + + x log + + 2 − x log =0
1 − x2 (1 − x2 )2 1 − x2 (1 − x2 )2 1 − x2 1 − x 1 − x2 1−x
Consider the general
P∞ theory for Legendre equation: (1 − x2 )y 00 − 2xy 0 + α(α + 1)y = 0. Let λ = α(α + 1). 1 − x2 , −2x, λ
analytic, so ∃ y = n=0 an xn .
∞
X ∞
X
=⇒ 2a2 + 3(2)a3 x + ((n + 2)(n + 1)an+2 − n(n − 1)an )xn + λan xn + λa1 x + λa0 =
n=2 n=2
∞
X ∞
X
=2 nan xn = 2 nan xn + 2a1 x
n=1 n=2
−λa0
a2 = (n(n + 1) − λ)an (n − α)(n + 1 + α)an
=⇒ 2 an+2 = =
(2 − λ)a1 (n + 2)(n + 1) (n + 2)(n + 1)
a3 =
6
If α = 1, λ = 2,
n−1 1
an+2 = an , a3 = 0, a2 = −a0 =⇒ a2m = (−a0 )
n+1 2m − 1
So that
∞
X x2m
y = −a0
m=0
2m − 1
Indeed,
∞ ∞
x X 2x2m+1 x2m
x 1+x X
log = = ,
2 1−x 2 m=0 2m + 1 m=1
2m − 1
x2m
P∞
so f (x) = − m=0 2m−1 .
Exercise 3.
a.
((x − a)(x − b)y 0 )0 − cy = 0 = ((At + B − a)(At + B − b)y 0 )0 − cy = 0
1 dt
Let x = At + B, c = α(α + 1), A = dx .
(At)2 + 2ABt + B 2 − (At + B)(b + a) + ab = A2 (t2 − 1)
b+a b−a
=⇒ 2AB − A(b + a) = 0 = A(2B − (b + a)) =⇒ B = ,A=
2 2
since
B 2 − B(b + a) + ab = −A2
(b + a)2 (b + a)2 −(b2 + 2ab + a2 ) −(b2 + −2ab + a2 ) −(b − a)2
=⇒ − + ab = = ab = = = −A2
4 2 4 4 4
67
b.
x(x − 1)y 00 + (2x − 1)y 0 − 2y = 0 = ((x2 − x)y 0 )0 − 2y = (x(x − 1)y 0 )0 − 2y
t+1
for x = 2
=⇒ ((t2 − 1)y 0 )0 − 2y = 0
Exercise 4. y 00 − 2xy 0 + 2αy = 0
∞
X 2(n − α)an
((n + 2)(n + 1)an+2 − 2nan + 2αan )xn =⇒ an+2 =
n=0
(n + 2)(n + 1)
For n = 2m
2(2m − 2 − α)a2m−2 −2(α − 2(m − 1))
a2m = = a2(m−1) =
(2m)(2m − 1) (2m)(2m − 1)
(−2)m (α − 2(m − 1))(α − 2(m − 2)) . . . α
= a0
(2m)!
For n = 2m + 1
2(2m − 1 − α)a2m−1 (−2)(α − (2m − 1))a2m−1
a2m+1 = = =
(2m + 1)(2m) (2m + 1)(2m)
(−2)m (α − (2m − 1))(α − (2m − 3)) . . . (α − 1)a1
=
(2m + 1)!
y = u1 + u2
∞ ∞
X (−2)m (α − (2m − 1))(α − (2m − 3)) . . . (α − 1) 2m+1 X (−2)m (α − 2(m − 1))(α − 2(m − 2)) . . . α 2m
= x + x
m=1
(2m + 1)! m=0
(2m + 2)!
u1 has a0 = 0, u2 has a1 = 0.
Since
u1 (0) = 0 u2 (0) = 1
u01 (0) =1 u02 (0) = 0
2(n−α)an
when α ∈ Z+ , then one of these u1 , u2 is a polynomial, since an+2 = (n+2)(n+1)
00 3 0 2
Exercise 5. For xy + (3 + x )y + 3x y = 0, assume an analytic expansion.
∞
X ∞
X
y= an xn = an−2 xn−2
n=0 n=2
X∞ ∞
X ∞
X
y0 = an nxn−1 = an−2 (n − 2)xn−3 = an+1 (n + 1)xn
n=1 n=3 n=0
X∞ ∞
X
y 00 = n(n − 1)an xn−2 = (n + 1)nan+1 xn−1
n=2 n=1
∞
X
2a2 x + 3(2)a3 x2 + 3a1 + 3a2 (2)x + 3a3 (3)x2 + 3a0 x2 + ((n + 1)nan+1 + 3an+1 (n + 1) + an−2 (n − 2) + 3an−2 ) xn = 0
n=3
a1 = 0
(n + 1)(n + 3)an+1 = −an−2 (n + 1)
8a2 = 0
=⇒ −an−2
−a an+1 =
(15a3 + 3a0 ) = 0 or a3 = n+3
5
−a3(j−1) (−1)2 a3(j−2) (−1)j a0
=⇒ a3j = = =
3j + 2 (3j + 2)(3j − 1) (3j + 2)(3j − 1) . . . (8)(5)
∞
X (−1)j
y = a0 {1 + x3j }
j=1
(3j + 2)(3j − 1) . . . (8)(5)
To obtain the solution with even-powered terms, consider first possible simple pole at 0 from the form of the differential
equation:
3
y 00 + + x2 y 0 + 3xy = 0
x
68
Then consider the following:
∞
X ∞
X
y= an xn−2 = an+1 xn−1
n=0 n=−1
X ∞
X
y0 = n = −1∞ (n − 1)an+1 xn−2 = (n + 2)an+4 xn+1
n=−4
∞
X
y 00 = (n + 2)(n + 1)an+4 xn+1
n=−4
Then for the first few terms,
(−2)(−3)a0 x−4 + (−1)(−2)a1 x−3 + 3(−2)a0 x−4 + 3(−1)a1 x−3 3a3 x−1 + (−2a0 )x−1 + 3a0 x−1 = 0
−a0
=⇒ a1 = 0, a3 =
3
∞
X (−1)j
y = x−2 a0 1 + x3j
j=1
3j j!
Exercise 6. x2 y 00 + x2 y 0 − (αx + 2)y = 0. αx+2
x2 analytic except at x = 0.
∞
X ∞
X
y= an xn = an−1 xn−1
n=0 n=1
X∞ ∞
X
y0 = nan xn−1 = (n − 1)an−1 xn−2
n=1 n=2
X∞
y 00 = n(n − 1)an xn−2
n=2
∞
X ∞
X ∞
X
(n(n − 1)an + (n − 1)an−1 )xn = αan−1 xn + 2an xn =
n=2 n=1 n=0
∞
X
= (αan−1 + 2an )xn + αa1 x + 2a1 x + 2a0
n=2
(α + 1 − n)an−1
(n − 2)(n + 1)an = (α + −n)an−1 or an =
(n − 2)(n + 1)
an α−(n−1)
an−1 = (n−2)(n+1) → 0 as n → ∞, so this power series converges ∀ x.
Also a0 = a1 = 0.
By recursion,
(α + 1 − n)(α + 2 − n)(α˙ − 2)
an = (3(2))a2
(n − 2)!(n + 1)!
Then,
∞ ∞
! !
X (α + 1 − n)(α + 2 − n)(α˙ − 2) X (α − n − 1)(α − n) . . . (α − 2) n+2
2
y = a2 x + 6xn = a2 2
x + 6x =
n=3
(n − 2)!(n + 1)! n=1
n!(n + 3)!
∞
!
2
X (α − n − 1)(α − n) . . . (α − 2) n
= a2 x 1+ 6x
n=1
n!(n + 3)!
Exercise 7. Leibniz’s formula for nth derivative of a product is the following: if h(x) = f (x)g(x), then
n
(n)
X n (k)
h (x) = f (x)g (n−k) (x)
k
k=0
69
a. For
∞
X
A(x) = an (x − x0 )n
n=0
X∞
B(x) = bn (x − x0 )n
n=0
and C(x) = A(X)B(x), then
n
X n (k)
C (n) (x) = A (x)B (n−k) (x)
k
k=0
b. Given that
A(k) (x0 ) = k!ak
B (n−k) (x0 ) = (n − k)!bn−k
Then
n n
X n X
C (n) (x0 ) = k!ak (n − k)!bn−k = n! ak bn−k
k
k=0 k=0
n
X
C (n) (x0 ) = n!cn so cn = a − kbn−k
k=0
Exercise 8.
1 dn 2
a. By Rodrigues’ formula, Pn (x) = − 1)n . (x2 − 1) = (x − 1)(x + 1). By Leibniz’s formula,
2n n! dxn (x
n k
dn n n
X n d dn−k
n
(x − 1) (x + 1) = k
(x − 1)n n−k (x + 1)n
dx k dx dx
k=0
n−1
1 1 X dk dn−k (x + 1)n
Pn (x) = (x + 1)n + n (x − 1)n n−k (x + 1)n = + (x − 1)Qn (x)
2n 2 n! dx k dx 2n
k=0
Qn (x) is a polynomial.
1 n dn
b. Pn (1) = 1. Pn (−1) = 0 + 2n n! (−2) n! = (−1)n where we considered when k = 0, for dxn (x + 1)n = n!.
c.
Exercise 9.
a. Now (1 − x2 )y 00 + −2xy 0 + α(α + 1)y = 0, or ((1 − x2 )y 0 )0 = −α(α + 1)y.
0 0
− m(m + 1)Pm = ((1 − x2 )Pm )
0 0 0
− m(m + 1)Pm Pn = ((1 − x2 )Pm ) Pn = ((1 − x2 )Pm Pn )0 − (1 − x2 )Pm
0
Pn0
n(n + 1)Pn Pm = −((1 − x2 )Pn0 Pm )0 + (1 − x2 )Pn0 Pm
0
0
=⇒ ((1 − x2 )(Pn Pm − Pn0 Pm ))0 = (n(n + 1) − m(m + 1))Pn Pm
R1
b. If n 6= m, −1
Pn Pm = 0
Exercise 10.
a. f (x) = (x2 − 1)n = (x − 1)n (x + 1)n . Using Leibniz’s rule again,
n−1
X dk n d
n−1+k
f (n−1) = (x − 1) (x + 1)n
dxk dxn−1+k
k=0
(n−1) (n−1)
For f (1) = 0, f (−1) = 0. Then
Z 1 1 Z 1 Z 1
(n) (n) (n−1) (n) (n−1) (n+1)
f f =f f − f f =− f (n−1) f (n+1)
−1 −1 −1 −1
Now
2n
d2n 2 2n dk 2n−k
n d
X
( n
f 2n)(x) = 2n (x − 1) = (x − 1) (x + 1)n = (2n)!
dx k dxk dx2n−k
k=0
for the k = n term.
Z 1 Z 1 Z 1 Z 1
f (n) f (n) = f (2n) f (0) (−1)n = (2n)! (1 − x2 )n dx = 2(2n)! (1 − x2 )n dx
−1 −1 −1 0
70
n
b. Now Pn (x) = 2n1n! dx
d 2 n
n (x − 1) .
Z 1 Z 1 Z 1 Z π/2
1 1 2(2n)!
(Pn (x))2 dx = 2n f (n) (n)
f = 2(2n)! (1 − x 2 n
) dx = sin2n+1 tdt =
−1 2 (n!)2 −1 22n (n!)2 0 22n (n!)2 0
2(2n)! (2n)!! 2(2n)!2n n!2n+1 22 (n + 1)
= 2n 2
= 2n 2
(n + 1)! = =
2 (n!) (2n + 1)!! 2 (n!) (2n + 2)! (2n + 2)(2n + 1)
2
=
2n + 1
6.24 E XERCISES - T HE METHOD OF F ROBENIUS , T HE B ESSEL EQUATION
Exercise 1.
(a) Given g(x) = x1/2 f (x), x2 y 00 + xy 0 + (x2 − α2 )y = 0 that f must satisfy,
1
g 0 = x−1/2 f + x1/2 f 0
2 , we want g to satisfy y 00 + 1 + 1−4α
2
y = 0.
−1 −3/2 4x2
y 00 = x f + x−1/2 f 0 + x1/2 f 00
4
Now
f0 α2
f 00 + + 1− 2 f =0
x x
2
−1 α
g 00 = −3/2 f + x1/2 − 1 f
4x x2
2 2
α 1 4α − 1
g 00 + g = x1/2 − f = g
x2 4x2 4x2
(b)
p P∞ (−1)n x 2n
(c) Jp (x) = x2
n=0 n!Γ(n+1+α) 2 . Also,
Γ n + 1 + 21 = n + 12 n − 1 + 21 . . . 1 + 12 12 Γ 12
1/2 X∞
2 (−1)n x 2n+1
J1/2 (x) = =
n!(n + 2 )(n − 1 + 12 ) . . . (1 + 21 )( 12 )Γ
1 1
x n=0 2
2
1/2 X ∞ 1/2
2 (−1)n 2n+1 2
= x = sin x
πx n=0
(2n + 1)! πx
Now for α = −1 2 , consider
Γ(n + 1 − 1/2) = Γ n + 12 = n + 12 − 1 n − 2 + 12 . . . 2 + 12 1 + 12 12 Γ 12
1/2 X ∞ x 2n 2 1/2 X∞ 1/2
2 (−1)n (−1)n x2n 2
J−1/2 (x) = 1
= = cos x
x n=0
n!Γ n + 2 2 x n=0
(2n)!Γ(1/2) πx
8.3 E XERCISES - F UNCTIONS FROM Rn TO Rm . S CALAR AND VECTOR FIELDS , O PEN BALLS AND O PEN SETS
Exercise 1. Let f be a scalar field defined on a set S and let c be a given real number. The set of all points x in S such that
f (x) = c is called a level set of f .
See sketch.
Exercise 2. Let S be the set of all points (x, y) in the plane satisfying the given inequalities.
See sketch.
Exercise 3. Proofs are hard! I read the examples at the end of Section 8.2, particularly the example on the 2-dim. Cartesian
product: it helps. In fact, we’ll review it right now.
A1 , A2 ⊆ R1
A1 × A2 = {(a1 , a2 )|a1 ∈ A1 , A2 ∈ A2 }
If A1 , A2 are open subsets of R1 ,
Choose any a ∈ A1 × A2
Want: a is an int. pt. of A1 × A2
71
Since
A1 , A2 open in R1 , ∃ B(a1 ; r1 ), ∃ B(a2 ; r2 )
Let r = min{r1 , r2 }
Want: B(a; r) ⊆ A1 × A2
If (x1 , x2 ) = x ∈ B(a; r),
kx − ak < r, so |x1 − a1 | < r, |x2 − a2 | < r2 ,
x1 ∈ B(a1 ; r1 ) x1 ∈ A1
then =⇒
x2 ∈ B(a2 ; r2 ) x 2 ∈ A2
We then get what we want: (x1 , x2 ) ∈ A1 × A2 so that any x ∈ B(a; r) belongs in S, which means, by def., that B(a; r) ⊆ S
Onward with the problem:
Let S be the set of all points (x, y, z) in 3-space.
(1) z 2 − x2 − y 2 − 1 > 0
(2) |x| < 1, |y| < 1, and |z| < 1 Consider a ∈ S We must use the fact that an open rectangular box is a basic open set.
Let a ∈(S, a = (a1 , a2 , a3 )
1 − ai if ai ≥ 0 Q3
Let ρi = and Ra = i=1 (ai − ρi , ai + ρi )
| − 1 − ai | if ai < 0
Consider b ∈ Ra .
If ai ≥ 0,
if bi ≥ ai , bi − ai < 1 − ai or bi < 1
if bi > 0, ai − bi > 0 or 1 > ai > bi
if bi < ai ,
if bi < 0, −bi < ai − bi < 1 − ai < 1
If ai < 0,
if bi > 0, bi − ai < 1 + ai or bi < 1 + 2ai < 1
if bi > ai , if bi < ai , ai − bi < 1 + ai or −bi < ai − bi <
if bi < 0, −bi < −a < 1 (since|ai | < 1)
1 + ai < 1.
Then, |bi | < 1 for each and every possible case. Then Ra ⊆ S, so ∀ a ∈ S is an int. pt. (since ∀ a, ∃ open
rectangle Ra , that is completely contained in S). Then S is open.
(3) x + y + z < 1
(4) |x| ≤ 1, |y| < 1, and |z| < 1
Consider a0 = (1, a2 , a3 ), 1 > a2 , a3 > 0.
a0 ∈ S, but for B(a0 , 1/2), (5/4, a2 , a3 ) ∈ B(a0 , 1/2) and (5/4, a2 , a3 ) ∈
/ S. So S is not open.
Or. . .
An open set has every element to be interior to it (definition).
An interior pt. is a pt. s.t. ∃ some open basic set containing the pt. and is a subset of the set.
We must show @ any open basic set for a pt. in this set.
Since every open rectangle contains an open ball and every open ball contains an open rectangle, we only need to
consider open rectangles.
Consider (1, y0 , z0 ) ∈ S.
Consider open rectangle containing 1. I claim that at best, 1 − n1 , 1 + n1 , n ∈ Z+ , since by Archimedes property
of real numbers,
Theorem 1 (Apostol’s Archimedes property of real numbers, pp. 26, Thm. 1.30, Vol. 1). If x > 0 and if y is an
arbitrary real number, ∃ n ∈ Z+ s.t. nx > y. We want i ∈ (ai , bi ) i.e. ai < 1 < bi
72
Consider open interval containing 1; 1 ∈ (ai , bi ). Then ai < 1 < bi . But (ai , bi ) always contains pts. not
belonging to Si : nx < y (existence of n guaranteed by Archimedes prop. of reals thm.).
1
nbi > (n + 1) =⇒ bi > 1 + >1
n
Then @ open interval containing 1 completely contained in S. (1, y0 , z0 ) is not an interior pt. S is not open.
(5) x + y + z < 1 and x > 0, y > 0, z > 0
Q3x ∈ S. Then x + y + z < 1.
Consider
Consider j=1 (xj − δj , xj + δj ).
3
X 3
X 3
X
xj + δ j = xj + δj
j=1 j=1 j=1
3
X
0 < x + y + z < 1 so let 1 − (xj ) = (x) > 0
j=1
3
X
We can choose δj = δ s.t. (x) > δ > 0
j=1
Furthermore, by Archimedes axiom, we can choose δj > 0 s.t. xj − δj > 0
Q3
=⇒ ∀ x ∈ S, we can construct an open rectangle j=1 (xj − δj , xj + δj ) = R(x) s.t. R(x) ⊆ S.
(6) x2 + 4y 2 + 4z 2 − 2x + 16y + 40z + 113 < 0
x2 + 4y 2 + 4z 2 − 2x + 16y + 40z + 113 = (x − 1)2 − 1 + 4(y + 2)2 − 16 + 4(z + 5)2 − 100 + 113 =
(x − 1)2
= (x − 1)2 + 4(y + 2)2 + 4(z + 5)2 − 4 < 0 =⇒ + (y + 2)2 + (z + 5)2 < 1
22
Thus, S is, by definition, a basic open set, a basic open sphere. By theorem, a basic open sphere is an open set.
Exercise 4.
(1) A is an open set in n-space and x ∈ A. Given A is open, A − {x} ⊂ A.
Consider a ∈ A − {x}. Then a ∈ A so, ∃Ba (a, ra ) ⊆ A.
If x ∈
/ Ba (a, ra ), we’re done.
If x ∈ Ba (a, ra ), consider rax = ka − xk
∀ xa ∈ Ba (a, rax ), kxa − ak < rax also kxa − ak < rax < ra , so xa ∈ A
=⇒ Ba (a, rax ) ⊆ A and Ba (a, rax ) ⊆ A − {x} since we constructed Ba s.t. x ∈
/ Ba
(2) A open. Let B have endpoints b1 , b2
A − {b1 } open. A − {b1 } − {b2 } = A0 open.
A0 − intB = A − B open since open set minus an open set is open, since union of 2 open sets is an open set.
We could also directly say, since we’re dealing with intervals in one-dimension, A = (aA , bA ), B = [aB , bB ].
bB < bA
.
aA < aB
B is a closed subinterval
S of A.
A − B =S(aA , aB ) (bB , bA )
(3) ∀ x ∈ A B, x ∈ A or x ∈ B. Since A, B are open, x is int. S to A, or int. to B.
Explicitly, if ∃ B(x, r) S
⊆ A, B thenSB(x, r) ⊆ A, B ⊆ A B.
Then x is interior
T to A B. =⇒ A B open.
∀ x ∈ A B, x ∈ A and x ∈ B.
Since x ∈ A and x ∈ B, ∃ B(x, rA ) ⊆ A; B(x, rB ) ⊆ B or
(x − rA , x + rA ) ⊆ A ; (x − rB , x + rB ) ⊆ B
Let rm = min (rA , rB ) T
So then (x − rm , x + rm ) ⊆ A and (x − rm , x + rm ) ⊆ B =⇒ (x − rm , x + rm ) ⊆ A B
(4) R1 is open (since ∀ B(x; r) ⊆ R1 )
A = [aA , bA ] is a closed interval.
Let R1 − A = R−
∀ x ∈ R− , x ∈ R1 so B(x; r) ⊆ R1 .
Suppose x ∈ R− , so x > bA or x < aA (otherwise x ∈ A).
73
If x > bA , then let r1 = x − bA ,
(x − r1 , x + r1 ) ⊆ R− since (x − r1 , x + r1 ) ⊆ R and ∀ x1 ∈ (x − r1 , x + r1 ), x1 > bA
If x < aA , then let r1 = aA − x
(x − r1 , x + r1 ) ⊆ R− since (x − r1 , x + r1 ) ⊆ R and ∀ x1 ∈ (x − r1 , x + r1 ), x1 < aA
Exercise 5. Prove the following properties of open sets in Rn
(1) The empty set ∅ is open.
Let a ∈ ∅
Then B(a; r) ⊆ ∅ since there are no a ∈ ∅, =⇒ a is an interior pt. of ∅.
Or . . .
Consider a ∈ ∅
Consider B(a; r) = ∅. Then B(a; r) ⊆ ∅. So ∅ is open.
(2) Rn is open.
Consider a ∈ Rn . Consider B(a; r). So for x ∈ B(a; r), then kx − ak < r.
=⇒ |xj − aj | < rj
xj ∈ R ∀ xj s.t. |xj − aj | < rj defines an open interval on R1 , and so by induction, the Cartesian product of n
open intervals is an open n-ball. So Rn is open.
Or . . .
Consider a ∈ Rn .
Consider B(a; r). Since ∀ y ∈ B(a; r), y ∈ Rn . B(a; r) ⊆ Rn . Rn open.
(3) Consider {WjS }, collection of open sets.
Consider y ∈ j Wj . Then y ∈ Wj for some j. Since Wj open, ∃ B(y; ρ) ⊆ Wj .
S S S
Wj ⊆ j Wj , so B(y, ρ) ⊆ j Wj . j Wj is open.
(4) Consider {WjT |j = 1, . . . , n}, finite collection of open sets.
n
Consider y ∈ j=1 Wj .
y ∈ Wi , ∀ i = 1, . . . , n. Then since
Tn ∀ i, Wi open, ∃ Bi (y, ρi ) ⊆ W
Ti n. Tn
By Thm., ∃ openset B(y, ρ) ⊆ i=1 Bi (y, ρi ). Then B(y, ρ) ⊆ i=1 Bi (y, ρi ) = i=1 Wi
−1 1
(5) Let WT
k = k , k ;k ≥1
Then k Wk = {0}, which is not open.
Exercise 7.
(1) (A {x})c = Ac (R − x)
S S
Ac open. Rn − x open. (sinceS{x} is closed). Then, by thm., the intersection of these two open sets, Ac (R − x) is
T
open. Then, by definition, A {x} is closed.
(2) R − [a, b] = [a, b]c .
Consider y ∈ R − [a, b]
If y > b, then
y − b > 0, so ∃ N ∈ Z+ s.t. y − b > N1 . y > N1 + b (Archimedes prop. of real numbers).
y ∈ N1 + b, y + 1 is open and N1 + b, y + 1 ⊆ R − [a, b]
If y < a, then
a − y > 0, so ∃ N ∈ Z+ s.t. a − y > N1 or a − N1 > y
y ∈ y − 1, a − N1 ⊆ R − [a, b]
thenSR − [a, b] open.
T c[a, b] closed (by definition).
c c
B . A, B closed, so Ac , B c open. Ac B c , intersection of 2 open sets, is open.
T
(3) (A B) S = A
then A B closed.
(A B)c = Ac B c . Ac , B c open. Ac B c open. Then (A B) closed.
T T S T
Exercise 8.
(1) ∅c = Rn and Rn open. ∅ closed.
n c
(2) T ) =c ∅ and
(R S ∅ open. Rn closed.
c c
S c T
(3) (Si Ai ) = iT Ai . Ai open, so i ATi open. i Ai closed.
n n n Sn
(4) ( i=1 Ai )c = i=1 Aci . Aci open, so i=1 Aci open. i=1 Ai closed.
S∞ +
S ∞ c S ∞
(5) i=1 {i} = Z is closed since ( i=1 {i}) = i=1 (i, i + 1) is open.
74
Exercise 9. Let S be a subset of Rn
(1) Prove that both intS and extS are open sets.
Want: intS is open, i.e. ∀ a ∈ intS, ∃ B(a; r) ⊆ intS i.e.
∀ x1 ∈ B(a; r), x1 ∈ intS
x1 ∈ intS if ∃ B(x1 ; r1 ) ⊆ S
Consider a ∈ intS, then ∃ B(a; r) ⊆ S
Consider x1 ∈ B(a; r). If kx1 − ak < r consider ∀ x2 s.t. kx2 − x1 k < kx1 − ak = r1 < r.
Then x2 ∈ B(a; r), so B(x1 , r1 ) ⊆ B(a; r) ⊆ S
=⇒ ∀ x1 ∈ intS for x1 ∈ B(a; r), so B(a; r) ⊆ intS.
Want: extS is open, i.e. ∀ a ∈ extS, ∃ B(a; r) ⊆ extS i.e.
∀ x1 ∈ B(a; r), x1 ∈ extS
x1 ∈ extS if ∃ B(x1 ; r1 ) s.t. ∀ x2 ∈ B(x1 , r1 ), x2 ∈
/ S.
Consider a ∈ extS, then ∃ B(a; r) s.t. ∀ x1 ∈ B(a, r), x1 ∈ /S
Consider x1 ∈ B(a; r). If kx1 − ak < r consider ∀ x2 s.t. kx2 − x1 k < kx1 − ak = r1 < r.
Then x2 ∈ B(a; r), so x2 ∈/ S. so then ∃ B(x1 , r1 ) s.t. ∀ x2 ∈ B(x1 , r1 ), x2 ∈
/S
=⇒ ∀ x1 ∈ extS, so B(a; r) ⊆ extS. extS open.
(2) Prove that Rn = (intS) (extS) ∂S, a union of disjoint sets, and use this to deduce that boundary ∂S is always
S S
a closed set.
Suppose ae ∈ extS. Then ∃ B(ae , r) s.t. ∀ xe ∈ B(ae , r), xe ∈ / S.
Then ∀ R > 0, B(ae , R) will contain xeR ∈ B(ae , R) s.t. xeR ∈ / S. (all open n-balls will either contain B(ae , r) or
be a part of B(ae , r)). So @ B(ae , R) s.t. B(ae , r) ⊆ S. ae ∈
/ intS
If ain ∈ intS, suppose ain ∈ extS. Then ain ∈
/ intS. Contradiction. ain ∈
/ extS.
intS, extS are open and disjoint.
abd is not interior to S, so abd ∈
/ intS
Suppose abd ∈ ∂S.
abd is not exterior to S, so abd ∈
/ extS
Let x ∈ Rn . Consider B(x, r0 ). If B(x, r0 ) ⊆ S, then x ∈ intS. If ∀ x1 ∈ B(x, r0 ), x1 ∈
/ S. Then x ∈ extS.
c
Otherwise, B(x, r0 ) may contain x1a ∈ S and x1b ∈ S . Then x is neither interior or exterior to S. So x ∈ ∂S
∂S, Rn ⊆ intS
S S S S
=⇒ x ∈ intS extS ∂S. extS
Since ∀ x ∈ intS extS ∂S, x ∈ R , then intS extS ∂S ⊆ Rn .
n
S S S S
=⇒ Rn = intS extS ∂S a union of disjoint sets.
S S
(intS extS)c = ∂S, intS extS is open so ∂S is closed, since its complement is open.
S S
Exercise 10. Want: x = boundary pt. of S = b.p. of S, neither interior nor exterior to S.
∀ B(x), ∃ ai ∈ B(x), s.t. ai ∈ intS ⊆ S. Then x cannot be an exterior pt.
∀ B(x), ∃ ae ∈ B(x), s.t. ae ∈ extS. Then ae ∈ / S and so x is not interior, by definition. x is a boundary pt.
Exercise 11. Rn − S = S c .
Let x ∈ intS c . Then ∃ open V s.t. x ∈ V and V ⊆ S c .
Then ∀ x1 ∈ V , x1 ∈/ S, so x is an exterior pt. to S. x ∈ extS, so intS c ⊆ extS.
Let x ∈ extS. Then ∃ B(x) s.t. B(x) ⊆ S c . By def., x ∈ intS c .
extS ⊆ intS c
=⇒ extS = intS c
Exercise 12. Suppose S closed. Let y be a boundary pt. of S.
Suppose y ∈/ S. Then y ∈ S c , S c open.
So by def. of open set, ∃ U s.t. y ∈ U and U ⊆ S c . But y is then an exterior pt., contradicting the definition of a boundary pt.
for y.
75
S
Then y ∈ S, so that S = intS ∂S
S
Suppose intS ∂S = S
Consider any z ∈ S c .
Then z has to be either a boundary pt. of S c or interior pt. of S c .
extS = intS c
z cannot be a boundary pt. of S c (we already showed that ), because then z ∈ ∂S and hence belong to S.
extS c = intS
Then z is interior to S c . S c = intS c , so S c open. S closed.
8.5 E XERCISES - L IMITS AND CONTINUITY
Exercise 1.
(1) f (x, y) is continuous ∀ (x, y) ∈ R2
(2) (x, y) 6= (0, 0)
(3) y 6= 0
2
(4) y 6= 0, xy 6= π2 + πk, k ∈ Z
(5) x 6= 0
(6) (x, y) 6= (0, 0)
x+y π
(7) 1−xy 6= 2 + πk, k ∈ Z, xy 6= 1
(8) (x, y) 6= (0, 0)
(9) f = exp (y 2 ln x), x 6= 0
(10) y 6= 0, xy ≥ 0
Exercise 2. limx→a f (x, y), limy→b f (x, y) exist, so limx→a f (x, y) = f (a, y) and limy→b f (x, y) = f (x, b).
Since limx→x0 f (x) = f (x0 ), where x0 = (a, b), which is equivalent to saying
∀ > 0, ∃ δ > 0 s.t. |f (x) − L| < if kx − x0 k < δ
then
Consider 2 > 0. ∃ δy > 0 s.t. |f (x, y) − f (x, b)| < 2 if |y − b| < δy (since limy→b f (x, y) exists).
Consider 2 > 0. ∃ δxy > 0 s.t. |f (x, y) − L| < 2 if k(x, y) − (a, b)k < δxy (since limx→x0 f (x) = L).
|f (x, b) − L| = |f (x, b) − f (x, y) + f (x, y) − L| < |f (x, y) − f (x, b)| + |f (x, y) − L| < + = whenever
p 2 2
|y − b| < δy and (x − a)2 + (y − b)2 < δxy
then |x − a| < δxy
So ∀ > 0, ∃ δxy = δx () s.t. |f (x, b) − L| < whenever |x − a| < δx ().
We just proved limx→a limy→b f (x, y) = limx→a f (x, b) = L
Similarly, we get the same result for limy tob f (a, y). Thus, limx→a limy→b f (x, y) = limy→b limx→a f (x, y) = L whenever
limx→x0 f (x) = L
(x−y)
Exercise 3. f (x, y) = x+y
lim f = 1
y→0
lim f = −1
x→0
x2 y 2
Exercise 4. f (x, y) = x2 y 2 +(x−y)2
lim f = 0
x→0 x4
but if y = x, f = =1
lim f = 0 x4 + 0
y→0
1
Exercise 5. 0 < x sin y <x x → 0, so by squeeze principle, x sin y1 → 0.
→ limx→0 f = 0
limy→0 f undefined, since
Consider |y| < n1 or n < |y| 1
For y > 0, sin y1 > sin n
For y < 0, | sin y1 | = sin −1
y > sin n
1 1
∀ δ = n , ∃ = (δ) = sin (1/δ) s.t. | sin 1/y| > if |y| < n
Then limy→0 limx→0 f = 0 6= limx→0 limy→0 f
76
x2 −y 2
Exercise 6. (x, y) 6= (0, 0), let f (x, y) = x2 +y 2
2 2 2
If y = mx, f → xx2 (1−m ) 1−m
(1+m2 ) = 1+m2 . If y = 0, f = 1. If x = 0, f = −1, so there’s no way to define f (0, 0) to be single
valued.
Exercise 7. Consider y = kx; k ∈ R.
For k = 0, y = 0 and f (x, y) = 0, if y = 0
For k ≷ 0, x ≶ 0, y < 0, so f (x, y) = 0.
Consider the limit as x → 0. can be as small as you want.
Then we must have |x| < < |k|.
then x2 < kx
Thus y = kx > x2 , so f (x, y) = 0 for any straight line through the origin.
Consider |x| < = 1
x4 < x2
4 2
So y = x < x → f (x, y) = 1
So, f is discontinuous at (0, 0). f (0, 0) depends upon path taken.
Exercise 8. Change to polar coordinates. Then
sin (x2 + y 2 ) sin r2
f (x, y) = = f (r, θ) =
(x2 + y 2 ) r2
2
Then regardless of what value of θ, limr→0 sinr2r = 1 .
Exercise 9. Let f be a scalar field continuous at an interior pt. a of a set S in Rn . f (a) 6= 0 (given).
Continuity of f at a means that
lim f (x) = f (a) =⇒ ∀ > 0, ∃ δ > 0 s.t. |f (x) − f (a)| < whenever kx − ak < δ
x→a
(
f (a) f (x) − f (a) < f (a)
2 if f (x) > f (a)
Let = 2 , ∃ δ = δ(; a) s.t. f (a) ,
−f (x) + f (a) < 2 if f (x) < f (a)
f (a) 3f (a)
so 2 < f (x) < 2 for ∀ x s.t. kx − ak < δ(; a).
δ(; a) defines a B(a) ⊆ Rn s.t. f (x) has the same sign as f (a).
8.9 E XERCISES - T HE DERIVATIVE OF A SCALAR FIELD WITH RESPECT TO A VECTOR , D IRECTIONAL DERIVATIVES
AND PARTIAL DERIVATIVES , PARTIAL DERIVATIVES OF HIGHER ORDER
Exercise 1. f (x) = a · x
f (x + hy) − f (x) a · (x + hy) − a · x
f 0 (x; y) = lim = lim = a·y
h→0 h h→0 h
Exercise 2. f (x) = kxk4
(1)
f (x + hy) − f (x) kx + hyk4 − kxk4
f 0 (x, y) = lim = lim = 4x2 (x · y)
h→0 h h→0 h
(2) n = 2
3 −2x 1
f 0 (2i + 3j; xi + yj) = 6 = 4(13)(2x + 3y) =⇒ = 2x + 3y =⇒ y = +
26 3 26
(3) n = 3
f 0 (i + 2j + 3k; xi + yj + zk) = 0 = 4(12 + 22 + 9)(x + 2y + 3z) =⇒ x + 2y + 3z = 0
Exercise 3.
f (x + hy) − f (x) (x + hy) · T (x + hy) − x · T (x) (x + hy) · (T (x) + hT (y)) − x · T (x)
f 0 (x, y) = lim = lim = lim =
h→0 h h→0 h h→0 h
hy · T (x) + hx · T (y) + h2 yT (y)
= lim = y · T (x) + x · T (y)
h→0 h
Exercise 4. f (x, y) = x2 + y 2 sin (xy)
∂x f = 2x + y 3 cos (xy)
∂y f = 2y sin (xy) + xy 2 cos (xy)
77
p
Exercise 5. f (x, y) = x2 + y 2
x y
∂x f = ; ∂y f =
f f
x
Exercise 6. f (x, y) = √
x2 +y 2
1 −x2
∂x f = + 2
f (x + y 2 )3/2
−xy
∂y f = 2
(x + y 2 )3/2
x+y
Exercise 7. f (x, y) = x−y , x 6= y
1 −(x + y) −2y
∂x f = + =
x−y (x − y)2 (x − y)2
1 −(x + y) 2x
∂y f = + 2
(−1) =
x−y (x − y) (x − y)2
Exercise 8. f (x) = a · x; a fixed.
∂xj f = aj
Pn Pn
Exercise 9. f (x) = i=1 j=1 aij xi xj , aij = aji
n X
X n n X
X n n
X n
X n
X
∂xk f (x) = aij δik xj + aij xi δjk = akj xj + aik xi = 2 akj xj
i=1 j=1 i=1 j=1 j=1 i=1 j=1
Exercise 10. f (x, y) = x4 + y 4 − 4x2 y 2
∂x f = 4x3 − 8xy 2 2
∂yx f = −16xy
∂y f = 4y 3 − 8x2 y 2
∂xy f = −16xy
Exercise 11. f (x, y) = log (x2 + y 2 )
2x −4xy
∂x f = ∂yx f =
x2 + y 2 (x2 + y 2 )2
2y −4xy
∂y f = 2 ∂xy f = 2
x + y2 (x + y 2 )2
1
Exercise 12. f (x, y) = y cos x2 ; y 6= 0
−2x sin x2 2x sin x2
∂x f = ∂yx f =
y y2
−1 2x sin x2
∂y f = 2 cos x2 ∂xy f =
y y2
Exercise 13. f (x, y) = tan (x2 /y); y 6= 0
2x
∂x f = sec (x2 /y)
y
−x2
∂y f = 2 sec (x2 /y)
y
Exercise 14. f (x, y) = arctan (y/x)
1 −y
∂x f =
1 + (y/x)2 x2
1 1
∂y f =
1 + (y/x)2 x
x+y
Exercise 15. f (x, y) = arctan 1−xy
78
1 − xy + xy + y 2 1 + y2
1 1 −(x + y)
∂x f = 2 = + ((−y) = =
1 − xy (1 − xy)2 1 + x2 y 2 + x2 + y 2 1 + x2 y 2 + x2 + y 2
x+y
1+ 1−xy
1 + x2
∂y f = (by label symmetry!)
1+ x2 y 2 + x2 + y 2
2
Exercise 16. f (x, y) = ey ln x x > 0
2
−1
∂x f = y 2 xy
2
∂y f = xy 2y ln x
p
Exercise 17. f (x, y) = arccos x/y; y 6= 0
−1
1/2 −1/2
∂x f = p √ =p
xy
1 − x/y xy − x2
√ √
−1 x(−1/2) (1/2) x/y
∂y f = p = √
1 − x/y y 3/2 y−x
2
Exercise 18. v(r, t) = tn e−r /4t
−1r n −r2 /4t −r n −r2 /4t
∂r v = t e = t e
2t 2t
−r3 n −r2 /4t
r2
r2 ∂r v = t e ∂t v = ntn−1 e−r
2
/4t
+ tn e−r
2
/4t
2t 4t2
−3r2 n −r2 /4t −r3 n −r
2
2
∂r (r ∂r v) = t e + t e−r /4t n = −3/2
2t 2t 2t
2 n
1 −3 2 r t 2
2
∂r (r2 ∂r v) = tn e−r /4t + 2 e−r /4t
r 2t 4t
∂2u
Exercise 19. z = u(x, y)eax+by ; ∂x∂y =0
∂x z = ∂x ueax+by + az
∂y z = (∂y u)eax+by + bz
2
∂xy z = a(∂y u)eax+by + b(∂x u)eax+by + baueax+by
2
∂xy z − ∂x z − ∂y z + z = a(∂y u)eax+by + b(∂x u) + abz − (∂x u)eax+by − az − (∂y u)eax+by − bz + z = 0
a = 1; b = 1
Exercise 20.
(1) f 0 (x, y) = 0 ∀ x ∈ B(a) ∀ y
Recall Thm. 8.4, Mean-value Thm. for derivatives of scalar fields. Assume ∃ f 0 (a + ty; y) ∀ t ∈ [0, 1]. Then ∃
some θ ∈ (0, 1) s.t.
f (a + y) − f (a) = f 0 (z; y), where z = a + θy
Proof. Let g(t) = f (a + ty)
Use one-dim. mean-value thm. to g on [0, 1].
g(1) − g(0) = g 0 (θ), θ ∈ (0, 1)
x=a+y
0
y=x; 0 ≤ |x0 | < r
∃ f 0 (x, y) = f 0 (a + ty; y) ∀ t ∈ [0, 1]
( since |ty| = t|y| < tr < r)
0
=⇒ f (a + ty; y) = 0 = f (a + y) − f (a)
f (a) = f (x), ∀ x ∈ B(a)
79
(2) Suppose we consider x = a + x0 where |x0 | < r and x0 k y.
y
f 0 (x, y) = f 0 (a + x0 , y) = f 0 (a + |x0 | , y) =
|y|
f (a + |x0 |ey , +h|y|ey ) − f (a + |x0 |ey ) f (a + |x0 |ey + (h|y|)ey ) − f (a + |x0 |ey )
= lim = lim =
h→0 h |y|h→0 h|y|/|y|
= |y|f 0 (a + |x0 |ey , ey )
∃ |y|f 0 (a + t|x0 |ey , ey ) ∀ t ∈ [0, 1], since f 0 (x, y) = 0 ∀ x ∈ B(a)
0 = |y|f 0 (a + t|x0 |ey , ey ) = f (a + |x0 |ey ) − f (a) =⇒ f (a + |x0 |ey ) = f (a)
f = f (a) = constant ∀ x ∈ B(a) s.t. x = a + key 0 ≤ |k| < r
Exercise 21.
(1) A set S in Rn is convex if ∀ a, b ∈ S,
ta + (1 − t)b ∈ S ∀ t ∈ [0, 1]
x1 = a + x01 kx1 − ak < r
Consider x1 , x2 ∈ b(a); and
x2 = a + x02 kx2 − ak < r
tx2 + (1 − t)x1 − a = t(a + x02 ) + (1 − t)(a + x01 ) − a = at + x02 t + a − at + x01 − tx01 − a =
= x01 (1 − t) + x02 t
ktx2 + (1 − t)x1 − ak = kx01 (1 − t) + x02 tk ≤ (1 − t)kx01 k + tkx02 k < (1 − t)r + tr = r
So tx2 + (1 − t)x1 ∈ S ∀ t ∈ [0, 1]. So an n-ball is a convex set.
(2) Consider x ∈ S. Then for some a, b ∈ S, k ∈ [0, 1], x = a + k(b − a).
choose y=b−a
f 0 (x; y) = f 0 (a + k(b − a), y) −−−−−−−−→ f 0 (a + k(b − a), b − a) exists
=⇒ f 0 (a + θ(b − a), b − a) = 0 = f (b) − f (a) =⇒ f (b) = f (a)
This must be true for all pairs of a, b ∈ S since x was arbitrarily chosen from S. f is constant on S.
Exercise 22.
(1)
f (a + hy) − f (a)
f 0 (a; y) = lim
h→0 h
f (a − hy) − f (a) f (a + (−h)y) − f (a)
f 0 (a, −y) = lim = − lim = −f 0 (a, y)
h→0 h −h→0 −h
so if f 0 (a, y) > 0, f 0 (a, −y) < 0
(2) f (x) = x · y because
f (x + hy) − f (x) x · y + hy 2 − x · y
f 0 (x; y) = lim = lim = y2 > 0
h→0 h h→0 h
8.14 E XERCISES - D IRECTIONAL DERIVATIVES AND CONTINUITY, T HE TOTAL DERIVATIVE , T HE GRADIENT OF A
SCALAR FIELD , A SUFFICIENT CONDITION FOR DIFFERENTIABILITY
Let’s review a number of important concepts with Rn fields. Differentiability must be redefined through a n − dim Taylor
expansion.
Definition 2 (Definition of a Differentiable Scalar Field).
Let f : S → R
Let a be an int. pt. of S.
Let B(a; r) s.t. B(a; r) ⊆ S
Let v s.t. kvk < r, so a + v ∈ B(a; r) Then
f diff. at a
if ∃ Ta , E s.t.
linear Ta : Rn → R
scalar E(a, v), E(a, v) → 0 as kvk → 0 and
(1) f (a + v) = f (a) + Ta (v) + kvkE(a, v)
The next theorem shows that if the total derivative exists, it is unique. It also tells us how to compute Ta (y), ∀ y ∈ Rn .
80
Theorem 2 (Uniqueness of total derivative). Assume f diff. at a with total derivative Ta
Then ∃ f 0 (a; y) ∀y ∈ Rn and
Ta (y) = f 0 (a; y)
Also,
X n
f 0 (a; y) = Dj f (a)yj for
j=1
y = (y1 , . . . , yj , . . . , yn )
Proof.
If y = 0, Ta (0) = 0 and f 0 (a; 0) = 0. Done.
Suppose y 6= 0
f (a + v) = f (a) + Ta (v) + kvkE(a, v) (since we assume f diff. )
v = hy
f (a + hy) − f (a) 1 khyk h→0
=⇒ = Ta (hy) + E(a, hy) −−−→ f 0 (a, y) = Ta (y) + 0
h h h
Now use linearity of Ta :
X X X X
Ta (y) = Ta (yj ej ) = yj Ta (ej ) = yj f 0 (a; ej ) = yj Dj f (a)
Pn was introduced, ∇f (a) = (∂1 f (a), . . . , ∂n f (a)) so that
Then the gradient
f 0 (a; y) = j=1 ∂j f (a)yj = ∇f (a) · y so then also
=⇒ f (a + v) = f (a) + ∇f (a) · v + kvkE(a; v)
Theorem 3 (Differentiability implies Continuity).
If a scalar field f is differentiable at a, then f is cont. at a
Proof. Since f is diff.
|f (a + v) − f (a)| = |∇f (a) · v + kvkE(a, v)|
By Cauchy-Schwarz inequality,
0 ≤ |f (a + v) − f (a)| ≤ k∇f (a)kkvk + kvk|E(a; v)|
As v → 0, |f (a + v) − f (a)| → 0 so f cont. at a.
If f is diff. at a, then all its partials exist (but the converse isn’t true).
existence of partials doesn’t necessarily imply f is diff.
2
e.g. f (x, y) = x2xy +y 4
Theorem 4 (Sufficient Condition for Differentiability).
Assume ∃ ∂1 f, . . . , ∂n f in some n-ball B(a) and are cont. at a. Then f diff. at a.
Proof.
Let λ = kvk; then v = λu, kuk = 1
Express f (a + v) − f (a) as a telescoping sum.
n
X
f (a + v) − f (a) = f (a + λu) − f (a) = (f (a + λvk ) − f (a + λvk−1 ))
k=1
v0 = 0
where {vk } s.t. . Then choose the vk ’s s.t.
vn = u
vk = vk−1 + uk ek
v1 = u1 e1 ; v2 = u1 e1 + u2 e2 , . . . , vn = u1 e1 + · · · + un en
f (a + λvk ) − f (a + λvk−1 ) = f (a + λvk−1 + λuk ek ) − f (a + λvk−1 ) =
= f (bk + λuk ek ) − f (bk )
81
bk , bk + λuk ek differ only by their kth component so apply the mean value theorem
=⇒ f (bk + λuk ek ) − f (bk ) = (λuk )∂k f (ck )
as bk → a, as λ → 0, so ck → a
n
X
=⇒ f (a + v) − f (a) = λ uk ∂k f (ck )
k=1
P
Now ∇f (a) · v = λ uk ∂k f (a). P
=⇒ f (a + v) − f (a) − ∇f (a) · v = λ uk (∂k f (ck ) − ∂k f (a)) = E(a, v)
ck → a as kvk → 0, and given ∂k f are cont., E(a, v) → 0 as kvk → 0.
By def. of diff., f is diff.
Exercise 1.
(1) f (x, y) = x2 + y 2 sin (xy)
∇f = (2x + y 2 cos (xy), 2y sin (xy) + y 2 x cos (xy))
(2) f (x, y) = ex cos y
∇f = (ex cos y, −ex sin y)
2 3 4
(3) f (x, y, z) = x y z
∇f = (2xy 3 z 4 , 3x2 y 2 z 4 , 4x2 y 3 z 3 )
(4) f (x, y, z) = x2 − y 2 + 2z 2
∇f = (2x, −2y, 4z)
2 2 2
(5) f (x, y, z) = log (x + 2y − 3z )
1
∇f = (2x, 4y, −6z)
f
z ln y
(6) f (x, y, z) = e(lnx)e
ez ln y
z
∇f = f , (ln x)ez ln y , (ln x)(ln y)ez ln y
x y
Exercise 2.
(1) f (x, y, z) = x2 + 2y 2 + 3z 2 at (1, 1, 0) in the direction of i − j + 2k.
f 0 (a, y) = ∇f (a) · y
(2) ∇f = (2x, 4y, 6z) ∇f (1, 1, 0) = (2, 4, 0). ∇f (a) · y = −2
Exercise 3. f (x, y) = 3x2 + y 2 ; x2 + y 2 = 1
(∇f ) · y = |∇f ||y| cos θ
p p p
|∇f | = 36x2 + 4(1 − x2 ) = 32x2 + 4 = 2 8x2 + 1; |∇f | maximized when x = ±1
(±1, 0), y k (±1, 0)
Exercise 4. (1, 2), +2 towards (2, 2); −2 towards (1, 1).
∇f (a) · y = ∇f (a) · (1, 0) = 2; ∇f (a) · (0, −1) = −2 =⇒ ∇f (a) = 2(1, 1)
(4, 6) − (1, 2) 14
∇f (a) · = ∇f (a) · (3, 4)/5 =
5 5
Exercise 5. a, b, c s.t. f (x, y, z) = axy 2 + byz + cz 2 x3 , (1, 2, −1)
∇f = (ay + 3cz x , 2axy + bz, by + 2czx3 )
2 2 2
∇f (1, 2, −1) = (4a + 3c, 4a + −b, 2b − 2c)
∇f (1, 2 − 1) · ez = 2b − 2c = 64 =⇒ b − c = 32
Maximum value means ∇f only has components in the z-direction.
∂x f = 4a + 3c = 0
c = −8; b = 24; a=6
∂y f = 4a − b = 0
Exercise 6. f 0 (a, y) = 1; f 0 (a, z) = 2 where y = 2i + 3j, z = i + j
82
(∂x f )(2) + (∂y f )(3) = 1 ∂y f = −3
(∂x f )(1) + (∂y f )(1) = 2 ∂x f = 5
Exercise 7. Let f and g denote scalar fields that are differentiable on an open set S.
(1) X
∇f (a) = (∂j f )(a)ej
(∂j f )(a) = f 0 (aj ej )
if f const., f 0 (a; ej ) = 0
=⇒ ∇f (a) = 0
0
We can also do the following: if ∇f = 0, f (a; y) = ∇f (a) · y = 0, ∀ y. Then, from Exercise 20 of Sec. 8.9, f is
constant on this open set S.
If f is constant on S, f (a + v) = f (a) for f (a + v) = f (a) + Ta (v) + kvkE(a, v)
Ta (y) = −kykE(a, y)
E(a, y) → 0 as y → 0
By uniqueness of the total derivative, ∇f (a) = 0, ∀ a ∈ S.
(2) ∇ is a linear transformation. =⇒ ∇(f + g) = ∇f + ∇g
(3) ∇ is a linear transformation. =⇒ ∇(cf ) = c∇f
(4)
(f g)(a + v) − (f g)(a) = ∇(f g)(a) · v + Ef g (a; v) = f (a + v)g(a + v) − f (a)g(a) =
= f (a + v)g(a + v) − f (a)g(a + v) + f (a)g(a + v) − f (a)g(a) =
= g(a + v)(f (a + v) − f (a)) + f (a)(g(a + v) − g(a)) =
= g(a + v)((∇f )(a) + Ef (a; v)) + f (a)((∇g)(a) + Eg (a; v))
Let kvk → 0, so that (∇(f g))(a) = g(a)(∇f )(a) + f (a)(∇g)(a)
(5)
f f f (a + v) f (a) g(a)f (a + v) − g(a + v)f (a)
(a + v) − (a) = − = =
g g g(a + v) g(a) g(a)g(a + v)
g(a)f (a + v) − g(a)f (a) + g(a)f (a) − g(a + v)f (a)
= =
g(a)g(a + v)
g(a)((∇f )(a) · v + Ef (a; v)) − f (a)((∇g)(a) · v + Eg (a; v))
= =
g(a)g(a + v)
f
=∇ · v + Ef /g (a; v)
g
g(a)∇f (a) − f (a)∇g(a) f
Let kvk → 0 =⇒ =∇ (a)
g 2 (a) g
Exercise 8. In R3 , let r(x, y, z) = xi + yj + zk, and let r(x, y, z) = kr(x, y, z)k
p
(1) ∇r = ∇ x2 + y 2 + z 2 = √ 2 1 2 2 (x, y, z) = ~rr
x +y +z
(2) Use induction.
~r ~r
∇(r2 ) = r + r = 2~r
r r
~r
∇(r ) = 2~rr + r2 = 3r~r
3
r
~r
∇(rn+1 ) = nrn−2~rr + rn−1 = (n + 1)rn−1~r
r
(3) n = 0. ∇(1) = 0
1 −~r
∇(r−1 ) = ∇ p =
x2 + + y2 z2 r2
1
∇(r−2 ) = ∇ 2 = (−2)~rr−4
x + y2 + z2
Then ∇(rn+1 ) = (n+1)rn−1~r, where we reuse the induction step above, because no reference was made to whether
n was positive or negative.
So the formula is still valid when n is a negative integer (by induction).
83
(4) ∇f = ~r
∂x f = x ∂y f = y ∂z f = z
1 2 1 2 1 2
x + y + z =f
2 2 2
Exercise 9. Given n independent vectors, y1 , . . . , yn , then by Thm., y1 , . . . , yn for a basis for Rn .
f 0 (x, y) = ∇f (x) · y, so f 0 (x, y) is linear. Then ∀ y ∈ Rn , y =
P
aj yj and
X X
f 0 (x, y) = ∇f (x) · aj yj = aj ∇f (x) · yj = 0
Then from Exercise 20 of Sec. 8.9, f is constant on B(a).
Exercise 10.
(1) Consider x ∈ B(a), x = a + x0 .
f 0 (x; y) = f 0 (a + x0 , y) = ∇f (x) · y
Let y = x0 =⇒ f 0 (a + x0 ; x0 )
By mean value thm., f 0 (a + θx0 ; x0 ) = f (a + x0 ) − f (a)
f 0 (a + x0 ; x0 ) = ∇f (x) · x0 = 0 =⇒ f (a + x0 ) = f (a)
This must be true ∀ x0 s.t. |x0 | < r for B(a; r). Then f is constant on B(a).
(2)
f (a + hy) − f (a)
lim = f 0 (a; y) ≤ 0
h→0 h
f 0 (a; y) = ∇f (a) · y = |∇f (a)||y| cos θ ≤ 0
Consider when −π/2< θ < π/2, |y| = 6 0. Then |∇f (a)| = 0.
Exercise 11. Consider the following six statements about a scalar field f : S → R, where S ⊆ Rn and a ∈ intS.
(1) (a) f continuous at a
(b) f is differentiable at a
(c) ∃ f 0 (a, y) ∀ y ∈ Rn .
(d) All the first-order partial derivatives of f exist in a neighborhood of a and are continuous at a.
(e) ∇f (a) = 0
(f) f (x) = kx − ak for all x in Rn .
(b) imples (a),(c), because differentiability implies continuity and differentiability through the total derivative gave what
the directional derivative would be ∀ y ∈ Rn . (d) imples (a),(b),(c) because (d), by theorem, is a sufficient condition for
differentiability, and thus differentiability implies (a),(c). (e) doesn’t tell us anything because we need a scalar function
E(a; v) as well for differentiability. (f) is a continuous function, so (f) imples (a).
8.17 E XERCISES - A CHAIN RULE FOR DERIVATIVES OF SCALAR FIELDS , A PPLICATIONS TO GEOMETRY. L EVEL SETS .
TANGENT PLANES .
Exercise 1.
(1)
u = f (x, y) x = X(t)
u = F (t) y = Y (t)
∇f (r) · r0 (t) = (∂x f )x0 + (∂y f )y 0 = F 0 (t) = u0
(2)
F (t) = u = f (x, y) = f (r(t))
F 0 (t) = ∇f (r) · r0 (t) = (∂x f )x0 + (∂y f )y 0
∇f (r) = ∇f (r(t))
d d
∇f (r(t)) = (∂x f, ∂y f ) = ((∂xx 2
f )x0 + (∂yx2
f )y 0 , (∂xy
2
f )x0 + (∂yy
2
f )y 0 )
dt dt
d
F 00 (t) = ∇f (r(t)) · r0 (t) + ∇f (r) · r00 (t) =
dt
2
f )x02 + (∂yx
2 2
f x0 y 0 + (∂yy2
f )y 02 + (∂x f )x00 + ∂y f y 00
= (∂xx f ) + ∂xy
84
Exercise 2.
(1) f (x, y) = x2 + y 2 , X(t) = t, Y (t) = t2
∂x f = 2x X0 = 1
∂y f = 2y Y 0 = 2t
F 0 (t) = 2x(1) + 2y2t = 2t + 4t3
F 00 (t) = 2 + 12t2
(2) f (x, y) = exy cos (xy 2 ); X(t) = cos t, Y (t) = sin t
∂x f = yf + −exy sin (xy 2 )y 2 X 0 = −s = −y X 00 = −c = −x
∂y f = xf + −exy sin (xy 2 )2yx Y0 =c=x Y 00 = −s = −y
F 0 (t) = −y 2 f + exy sin (xy 2 )y 3 + x2 f + −exy sin (xy 2 )2yx2
2
This is the answer I got. Note that I tried it 2 ways: using the formula (∂xx f )x02 + (∂yx
2 2
f + ∂xy f )x0 y 0 + (∂yy
2
f )y 02 +
00 00 0 0 0
(∂x f )x + (∂y f )y , and second, taking our answer F (t) = G(t) and then applying ∂gx x + ∂gy y on it (which
seemed clever).
∂x g = 4xf + (2x2 − 1)(yf − y 2 exy sin (xy 2 )) + (3y 3 − 2y)yexy sin (xy 2 ) + (3y 3 − 2y)exy cos (xy 2 )y 2 =
= 4xf + 2x2 yf − yf − 2x2 y 2 exy sin (xy 2 ) + y 2 exy sin (xy 2 ) + 3y 4 exy sin (xy 2 ) − 2y 4 exy sin (xy 2 ) + 3y 5 f − 2y 3 f
= f (4x + 2x2 y − y + 3y 5 − 2y 3 ) + exy sin (xy 2 )(−2x2 y 2 + y 2 + y 4 )
∂y g = (2x2 − 1)(xf − 2yxexy sin (xy 2 )) + (9y 2 − 2)exy sin (xy 2 ) + (3y 3 − 2y)exy x sin (xy 2 ) + (3y 3 − 2y)exy cos (xy 2 )2yx =
= (2x3 − x + 6y 4 x − 4y 2 x)f + exy sin (xy 2 )(−2yx(2x2 − 1) + 9y 2 − 2 + 3y 3 x − 2yx)
∂x gx0 + ∂y gy 0 = (−12y 5 + 14y 3 − 4y + 7x − 9x3 )exy sin (xy 2 ) + f (9x6 − 11x4 + 3x2 − 4xy)
2
(1+ex ) 2 2
(3) f (x, y) = log 1+ey2 = log (1 + ex ) − log (1 + ey )
X(t) = et X0 = X X 00 = X
Y (t) = e−t Y 0 = −Y Y 00 = Y
1 2
∂x f = (2xex ) 2 2
1 + ex2 2x2 ex 2y 2 ey
2 F 0 (t) = 2 +
−2yey 1 + ex 1 + ey 2
∂y f =
1 + ey 2 ! !
2 2 2 2 2 2 2 2
(ex + 2x2 ex )(1 + ex ) − (2xex )(xex ) ex + 2x2 ex + e2x
∂xx f = (2) = (2)
(1 + ex2 )2 (1 + ex2 )2
2 2 2
!
ey + 2y 2 ey + e2y
∂yy f = (−2)
(1 + ey2 )2
∂xy f = 0
2
(∂xx f )x02 + ((∂yx
2 2
f + ∂xy f )x0 y 0 + (∂yy
2
f )y 02 + (∂xf )x00 + ∂yf y 00 =
2 2 2 2 2 2 2 2 2 2
ex + 2x2 ex + e2x 2 ey + 2y 2 ey + e2y 2 2xex (1 + ex ) −2yey (1 + ey )
= (2) (x ) + (−2) y + x + y=
(1 + ex2 )2 (1 + ey2 )2 (1 + ex2 )2 (1 + ey2 )2
2 2 2 2
4x2 ex (1 + x2 + ex ) −4y 2 ey (1 + y 2 + ey )
= +
(1 + ex2 )2 (1 + ey2 )2
Exercise 3.
(1)
∇f = (3, −5, 2)
N = (2x, 2y, 2z) = 2r
(2, 2, 1)
∇f · = (3, −5, 2) · (2, 2, 1)/3 = −2/3
3
It should be noted that the normal to a sphere is the position vector.
85
(2)
∇f = (2x, −2y, 0)
x2 + y 2 + z 2 = 4 is a sphere
(x, y, z) 2x2 − 2y 2
∇f · =
r r
(3) x2 + y 2 = 25 and 2x2 + 2(z 2 − x2 ) − z 2 = 25
∇f = (2x, 2y, −2z)
√
∓x √ ∓x , 0
!
(1, √25−x 2
, 0) 1, 25−x 2 25 − x2 ∓x
T = q =p = , ,0
x2
1 + 25−x 25/(25 − x2 ) 5 5
2
√ √ !
2x 25 − x2 −2 25 − x2 x
6 −2(4)3 24 −24
=⇒ ∇f · T = , ,0 = 4, ,0 = + +0 =0
5 5 5 5 5 5
Exercise 4.
(1) Find a vector V (x, y, z) normal to the surface
p
z = x2 + y 2 + (x2 + y 2 )3/2
at a general point (x, y, z) of the surface, (x, y, z) 6= (0, 0, 0)
p
0= x2 + y 2 + (x2 + y 2 )3/2 − z = f (r)
!
x 2 2 1/2 y 2 2 1/2
=⇒ ∇f = p + 3x(x + y ) ,p + 3y(x + y ) , −1
x2 + y 2 x2 + y 2
(2)
−1
∇f · ez = |∇f | cos θz =⇒ cos θz = q
(x+3x3 +3xy 2 )2 (y+3yx2 +3y 3 )2
x2 +y 2 + x2 +y 2 +1
−1
cos θz = p
(1 + 3(x2 + y 2 ))2 + 1
−1 x→0 −1
lim cos θz = p −−−→= √
y→0 (1 + 3(x2 ))2
+1 2
√ √
lim lim cos θz = −1 2
x→0 y→0
cos θz is differentiable at (x, y) = (0, 0) (we can observe that the partial derivatives exist and are continuous at
(0, 0)), so cos θz is continuous at (x, y) = (0, 0).
eu cos v = x u = u(x, y)
Exercise 5. u
e sin v = y v = v(x, y)
y
sin v = p
x2 + y 2 = e2u 1 + y2x2 y
=⇒ ln (x2 + y 2 ) = u =⇒ tan v =
2 2
ln (x + y ) = 2u 2 x x
cos v =
sqrtx2 + y 2
x y
∇U = ,
x2 + y 2 x2 + y 2
=⇒ ∇U · ∇V = 0
−y/x2
1/x
∇V = ,
1 + (y/x)2 1 + (y/x)2
p
Exercise 6. f (x, y) = |xy|
86
(1) r r
1/2 1 y 1 x
if x ≷ 0, y ≷ 0, f = (xy) ∂x f = ∂y f =
2 x 2 y
1/2 1/2
|y| 1 −1 x
if x > 0, y < 0, f = (x(−y))1/2 ∂x f = ∂y f =
x 2 2 |y|
1/2 1/2
−1 y 1 |x|
if x < 0, y > 0, f = (−xy)1/2 ∂x f = ∂y f =
2 |x| 2 y
(2) Does the surface z = f (x, y) have a tangent plane at the origin?
z = f (x, y)
g = f (x, y) − z =⇒ ∇g = (∂x f, ∂y f, −1)
For x = y,
1 1
∇g(0, 0, 0) = , , −1
2 2
But when approaching from the x or y axis, ∇g = (0, 0, −1). A tangent plane cannot be defined at the origin.
Exercise 7. Given surface z = xy, z = y0 x, y = y0 and z = x0 y, x = x0 intersect at (x0 , y0 , z0 ) and lie on the surface. We
want to show that the tangent plane to this surface at (x0 , y0 , z0 ) contains these 2 lines.
Note that the 2 lines could be reexpressed in vector form:
x(1, 0, y0 ) + (0, y0 , 0)
y(0, 1, x0 ) + (x0 , 0, 0)
Rewrite the surface equation so to get the gradient
0 = xy − z
∇f = (y, x, −1)
∇f (r0 ) = (y0 , x0 , −1)
∇f (r0 ) · (x, y, z) = y0 x + x0 y − z = x0 y0
(0, y0 , 0) ∈ S
∇f (r0 ) · (1, 0, y0 ) = ∇f (r0 ) · (0, 1, x0 ) = 0 and note that
(x0 , 0, 0) ∈ S
So indeed, the tangent plane contains these lines.
Exercise 8. xyz = a3 (x0 , y0 , z0 ), ∇f = (yz, xz, xy)
y0 z0 x + x0 z0 y + x0 y0 z = 3a3
Volume of the tetrahedron:
9a3
1 1 1 1 1
V = Bh = xy h = xyz = (3x0 )(3y0 )(3z0 ) =
3 3 2 6 6 2
Exercise 9. We want a pair of linear Cartesian equations for the line tangent to x2 +y 2 +2z 2 = 4, z = ex−y at pt. (1, 1, 1) = P1
We calculate the gradients for the 2 surfaces, so to get the normal to these surfaces.
∇f = (2x, 2y, 4z) ∇g = (ex−y , −ex−y , −1)
(1,1,1)
−−−−→ ∇f (1, 1, 1) = 2(1, 1, 2) ∇g(1, 1, 1) = (1, −1, −1)
We know the general form of the equation for the line with these normals, N , will be X · N = X · P1 . Then
x − y − z = −1
x + y + 2z = 4
Exercise 10. Find a constant c s.t. at any pt. of intersection, the corresponding tangent planes will be ⊥ to each other.
(x − c)2 + y 2 + z 2 = 3 ∇f = 2(x − c, y, z)
2 2 ∇g = 2(x, y − 1, z)
x + (y − 1) + z = 1
tangent planes are ⊥ to each other =⇒ ∇f · ∇g = x(x − c) + y(y − 1) + z 2 = x2 − xc + y 2 − y + z 2 = 0
=⇒ y = xc
87
We have the intersection condition, and so solving for the system of 2 linear equations, with y = xc,
√
c=± 3
Exercise 11. Without loss of generality, choose the origin to make the ellipse symmetrical and the major axis to lie on the x
axis.
R = (x, y)
p
F1 = (ae, 0) |R − F1 | = r1 = (x − ae)2 + y 2 R − F1 = (x − ae, y)
R − F2 = (x + ae, y)
p
F2 = (−ae, 0) |R − F2 | = r2 = (x + ae)2 + y 2
(x − ae, y) R − F1
∇|R − F1 | = p =
(x − ae)2 + y 2 r1
(x + ae, y) R − F2
∇|R − F2 | = p =
(x + ae)2 + y2 r2
r1 + r2 = K
x2
2 2
y = b 1 −
x2 y2 a2 dy ∓bx/a2 −b2 x
+ = 1 = =
a2 b2 a2 y
r p
x2 dx 1 − (x/a)2
y = ±b 1 − 2
a
0 −b2 x
R = (1, 2 )
a y
R0 · kR + R0 · (−F1 r2 − F2 r1 )
0 0 R − F1 R − F2
R · ∇(|R − F1 | + |R − F2 |) = R · + =
r1 r2 r1 r2
2
Consider R0 · kR + r2 (−F1 · R0 ) − r1 (R0 · F2 ) = K(x − ab 2 x) + r2 (−ae) − r1 (−ae)
r r
b 2 x2 b2 x2 p
r1,2 = (x ∓ ae)2 + b2 − 2 = x2 ∓ 2xae + a2 e2 + b2 − 2 = e2 x2 ∓ 2xae + a2 = a ∓ xe
a a
r2 − r1 = 2xe
r2 + r1 = 2a
=⇒ R · kR + r2 (−F1 · R ) − r1 (R0 · F2 ) = 2axe2 + (−ae)(2xe) = 0
0 0
Thus T · (∇r1 + r2 ) = 0. This means that T · ∇r1 = −T · ∇r2 . As we had shown above, ∇r1,2 is in the direction from the
respective foci to the arbitrary point (x, y) and both ∇r1 and ∇r2 are of length 1. Thus T · ∇r1 = −T · ∇r2 geometrically
says that the angle between ∇r1 and the tangent line is equal to the angle between ∇r2 and the tangent line.
Exercise 12. f = f (x, y, z)
∂z f = k0 z =⇒ f = k1 z 2 g(x, y) + h(x, y)
f (0, 0, a) = k1 a2 g(0, 0) + h(0, 0) = f (0, 0, −a)
8.22 E XERCISES - D ERIVATIVES OF VECTOR FIELDS , D IFFERENTIABILITY IMPLIES CONTINUITY, T HE CHAIN RULE
FOR DERIVATIVES OF VECTOR FIELDS , M ATRIX FORM OF THE CHAIN RULE
Exercise 1. Recall
n
X n
X
(Dh(a))jk = (Df (b))jl (Dg(a))lk = (∂k hj (a)) = (∂l fj (b))(∂k gl (a))
l=1 l=1
(1)
∂f ∂g
∂x f = ∂t F ∂x g =⇒ = F 0 (g(x, y))
∂x ∂x
∂f 0 ∂g
= F (g(x, y))
∂y ∂y
(2)
∂f
F (t) = esin t = − cos tesin t sin (x2 + y 2 )2x
∂x
g(x, y) = cos (x2 + y 2 ) ∂f
= − cos tesin t sin (x2 + y 2 )2y
∂y
88
Exercise 2. Given
x−y
u=
2 f (u, v) → F (x, y), F (x, y) = f (u(x, y), v(x, y))
x+y
v=
2
1 1
∂x F = ∂u f ∂x u + ∂v f ∂x v = ∂u f + ∂v f
2 2
1 1
∂y F = ∂u f ∂y u + ∂v f ∂y v = − ∂u f + ∂v f
2 2
Exercise 3. Given
x = X(s, t)
u = f (x, y) u = F (s, t) = f (x(s, t), y(s, t))
y = Y (s, t)
(1)
∂s F = ∂x f ∂s x + ∂y f ∂s y
∂t F = ∂x f ∂t x + ∂y f ∂t y
(2) To get to the second order partial derivatives, it seems that a direct application of the partial derivatives is needed:
there’s not a way to reformulate a matrix chain rule for second order partial derivatives, until maybe the Hessian
matrix.
2 2 2 2 2
∂ss f = ∂xs F ∂s x + ∂x f ∂ss x + ∂ys F ∂s y + ∂y f ∂ss y
∂s f = ∂x f ∂s x + ∂y f ∂s y
=⇒ ∂xs f = ∂xx f ∂s x + ∂x f ∂s ∂x x + ∂xy f ∂s y + ∂y f ∂s ∂x y =
= ∂xx f ∂s x + ∂xy f ∂s y
2 2 2 2 2 2 2
=⇒ ∂ss f = (∂xx f ∂s x + ∂xy f ∂s y)∂s x + ∂x f ∂ss x + (∂yx f ∂s x + ∂yy f ∂s y)∂s y + (∂y f )(∂ss y) =
2 2 2 2 2 2 2
= (∂xx f )(∂s x) + 2∂xy f ∂s y∂s x + ∂yy f (∂s y) + (∂x f )∂ss x + (∂y f )(∂ss y)
(3) By label symmetry:
2 2
∂tt f = (∂xx f )(∂t x)2 + 2∂xy
2 2
f ∂t y∂t x + ∂yy f (∂t y)2 + (∂x f )∂tt
2 2
x + (∂y f )(∂tt y)
2
Let’s calculate ∂st F
2 2 2 2 2
∂st F = ∂xs f ∂t x + ∂x f ∂st x + ∂ys f ∂t y + ∂y f ∂st y=
2 2 2 2 2 2
= (∂xx f ∂s x + ∂xy f ∂s y)∂t x + ∂x f ∂st x + (∂yx f ∂s x + ∂yy f ∂s y)∂t y + ∂y f ∂st y=
2 2 2 2 2
= ∂xx f ∂s x∂t x + ∂xy f (∂s y∂t x + ∂s x∂t y) + ∂yy f ∂s y∂t y + ∂x f ∂st x + ∂y f ∂st y
Exercise 4.
(1)
X(s, t) = s + t ∂s X = ∂t X = 1 ∂s F = ∂x f + t∂y f
Y (s, t) = st ∂s Y = t ∂t Y = s ∂t F = ∂x f + s∂y f
2 2 2 2
∂ss f = (∂xx f ) + 2∂xy f t + ∂yy f t2
2 2 2 2
∂tt f = (∂xx f ) + 2∂xy f s + ∂yy f s2
2 2 2 2
∂st f = ∂xx f + ∂xy f (t + s) + ∂yy f ts
(2)
X(s, t) = st Xs = t Ys = 1/t ∂s F = ∂x f t + ∂y f (1/t)
Y (x, t) = s/t Xt = s Yt = −s/t ∂t F = ∂x f s + ∂y f (−s/t2 )
2 2 1
∂ss F = (∂xx f )(t2 ) + 2∂xy
2
f 2
t + ∂yy f (1/t)2
t
2
2 2 2 2 −s 2 −s
∂tt F = (∂xx f )s + 2∂xy f s + ∂yy f
t t2
2 2 2 1 −s 2 1 −s −1
∂st F = ∂xx f ts + ∂xy f s+t + ∂yy f + ∂ x f + ∂ y f
t t2 t t2 t2
89
(3)
s−t
X(s, t) = xs = 1/2 Ys = 1/2 ∂s F = ∂x f 1/2 + ∂y f 1/2
2
s+t xt = −1/2 Yt = 1/2 ∂t F = ∂x f (−1/2) + ∂y f 1/2
Y (s, t) =
2
2 1 2 2 2 1 2 2 2 2
∂ss F = ((∂xx f + ∂xy f )1/2 + (∂yx f + ∂yy )1/2) = (∂xx f + ∂xy f + ∂yx f + ∂yy f)
2 4
2 1 2 2 2 2
∂tt F = (∂xx f − ∂xy f − ∂yx f + ∂yy f)
4
2 1 2 2 2 2
∂st F = (−∂xx f + ∂xy − ∂yx f + ∂yy f)
4
Exercise 5. You cannot interchange ∂x and ∂r , ∂x and ∂θ , etc.
∂r φ = ∂x f ∂r x + ∂y f ∂θ y = fx c + fy s
∂θ φ = ∂x f ∂θ x + ∂y f ∂θ y = −fx rs + fy rc
Notice that the above formulas give a prescription or algorithm for computing the ∂r or ∂θ of functions of x, y. Notice also
that fx , fy are each composite functions.
2
∂rθ φ = ∂r (−fx rs + fy rc) = −∂r fx rs − fx s + ∂r fy f c + fy c =
= −(fxx c + fyx s)rs − fx s + (fxy c + fyy s)rc + fy c =
= −fxx rcs − fyx rs2 + fxy rc2 + fyy rsc − fx s + fy c
2
∂θr φ = ∂θ (fx c + fy s) = ∂θ fx c − fx s + ∂θ fy s + fy c =
= (−fxx rs + fyx rc)c − fx s + (−fxy rs + fyy rc)s + fy c =
= −fxx rsc + fyx rc2 − fx s − fxy rs2 + fyy rcs + fy c
2
∂rr φ = ∂r (fx c + fy s) =
= ∂r fx c + ∂r fy s = (fxx c + fyx s)c + (fxy c + fyy s)s = fxx c2 + fyx sc + fxy cs + fyy s2
x = X(r, s, t)
Exercise 6. Given u = f (x, y, z), y = Y (r, s, t) u = F (r, s, t)
z = Z(r, s, t)
∂r F = ∂x F ∂r x + ∂y F ∂r y + ∂z F ∂r z
∂s F = ∂x F ∂s x + ∂y F ∂s y + ∂z F ∂s z
∂t F = ∂x F ∂t x + ∂y F ∂t y + ∂z F ∂t z
Exercise 7.
X(r, s, t) = r + s + t
(1) Given Y (r, s, t) = r + −2s + 3t
Z(r, s, t) = 2r + s + −t
∂r F = ∂x F + ∂y F + 2∂z F
∂s F = ∂x F + −2∂y F + ∂z F
∂t F = ∂x F + 3∂y F − ∂z F
X(r, s, t) = r2 + s2 + t2
(2) Given Y (r, s, t) = r2 − s2 − t2
Z(r, s, t) = r2 − s2 + t2
∂r F = 2r(∂x F + ∂y F + ∂z F )
∂s F = 2s(∂x F + −∂y F − ∂z F )
∂t F = 2t(∂x F − ∂y F + ∂z F )
90
x = X(s, t)
Exercise 8. u = f (x, y, z) y = Y (s, t) u = F (s, t)
z = Z(s, t)
∂s F = ∂x F ∂s X + ∂y F ∂s Y + ∂z F ∂s Z
∂t F = ∂x F ∂t X + ∂y F ∂t y + ∂z F ∂t Z
Exercise 9.
X(s, t) = s2 + t2
∂s F = 2s(∂x F + ∂y F ) + 2t∂z F
(1) Y (s, t) = s2 − t2
∂t F = 2t(∂x F − ∂y F ) + 2s∂z F
Z(s, t) = 2st
X(s, t) = s + t
∂s F = (∂x F + ∂y F ) + t∂z F
(2) Y (s, t) = s − t
∂t F = (∂x F − ∂y F ) + s∂z F
Z(s, t) = st
∂r F = ∂x F ∂r x + ∂y F ∂r y
x = X(r, s, t)
Exercise 10. Given u = f (x, y); u = F (r, s, t) =⇒ ∂s F = ∂x F ∂s x + ∂y F ∂s y
y = Y (r, s, t)
∂t F = ∂x F ∂t x + ∂y F ∂t y
Exercise 11.
∂r F = ∂x F
(1) Given X(r, s, t) = r + s, Y (r, s, t) = t =⇒ ∂s F = ∂x F
∂t F = ∂y F
∂r F = ∂x F + ∂y F (2r)
2 2 2
(2) Given X(r, s, t) = r + s + t, Y (r, s, t) = r + s + t =⇒ ∂s F = ∂x F + 2s∂y F
∂t F = ∂x F + 2t∂y F
1
∂x F
∂r F =
s
(3) Given X(r, s, t) = r/s, Y (r, s, t) = s/t =⇒ ∂s F = ∂x F (−r/s2 ) + ∂y F/t
∂t F = ∂y F (−s/t2 )
Exercise 12. h(x) = f (g(x)) g = (g1 , . . . , gn )
n
X n
X
∇h(a) =⇒ ∂k h(a) = ∂l f ∂k gl or ∇h(a) = ∂l f ∇gl
l=1 l=1
Exercise 13.
(1) f (x, y, z) = xi + yj + zk
∇fx (x) ∂x fx ∂y fx ∂z fx 1
Df (x) = ∇fy (x) = ∂x fy ∂ y fy ∂z fy = 1
∇fz (x) ∂ x fz ∂y fz ∂z fz 1
(2) f = (x + cx , y + cy , z + cz ) where cz , cy , cz are constants.
(3)
∇fx (x) ∂x fx ∂y fx ∂z fx p(x)
Df (x) = ∇fy (x) = ∂x fy ∂y fy ∂z fy = q(y)
∇fz (x) ∂x fz ∂y fz ∂z fz r(z)
Z Z Z
=⇒ f (x) = (( p(x)dx + x0 ), ( q(y)dy + y0 ), ( r(z)dz + z0 ))
where x0 , y0 , z0 are constants.
f (x, y) = (ex+2y , sin (y + 2x))
Exercise 14. Given f : R2 → R2 , g : R3 → R2
g(u, v, m) = ((u + 2v 2 + 3w3 ), (2v − u2 ))
ex+2y 2ex+2y 9w2
1 4v
(1) Df (x, y), Dg(u, v, w) =⇒ Df = Dg =
2 cos (y + 2x) cos (y + 2x) −2u 2 0
91
(2) h(u, v, w) = f (g(u, v, w))
2
+3w3 +2(2v−u2 )
f (g(u, v, w)) = (eu+2v , sin (2v − u2 + 2(u + 2v 2 + 3w3 ))) =
2
+3w3 +4v−2u2
= (eu+2v , sin (2v − u2 + 2u + 4v 2 + 6w3 ))
(3)
ex+2y 2ex+2y 4v 9w2
1
Dh = Df Dg = =
2 cos (y + 2x) cos (y + 2x) −2u 2 0
ex+2y (1 − 4u) ex+2y (4v + 4) 9w2 ex+2y
=
cos (y + 2x)(2 − 2u) cos (y + 2x)(8v + 2) 18w2 cos (y + 2x)
−3 0 9
Dh(1, −1, 1) =
0 −6 cos 9 18 cos 9
Exercise 15. Given
f = ((x2 + y + z), (2x + y + z 2 ))
g = (uv 2 w2 , w2 sin v, u2 ev )
2 2
2uvw2 2uv 2 w
v w
2x 1 1
(1) Df = Dg = 0 w2 cos v 2w sin v
2 1 2z
2uev u2 ev 0
(2)
h(u, v, w) = f [g(u, v, w)] = ((uv 2 w2 )2 + w2 sin v + u2 ev , 2uv 2 w2 + w2 sin v + u4 e2v ) =
= (u2 v 4 w4 + w2 sin v + u2 ev , 2uv 2 w2 + w2 sin v + u4 e2v )
(3)
v 2 w2 2uvw2 2uv 2 w
2x 1 1
Dh(u, 0, w) = 0 w2 cos v 2w sin v =
2 1 2z v
2ue u2 ev 0
2 2 v
4xuvw + w cos v + u2 ev
2 2
4xuv 2 w + 2w sin v w2 + u2
2xv w + 2ue 2u 0
= 2 2 v =
2v w + 4zue 4uvw2 + w2 cos v + 2zu2 ev 2
4uv w + 2w sin v 4(u ) w2 + 2u4
3
0
8.24 M ISCELLANEOUS EXERCISES - S UFFICIENT CONDITIONS FOR THE EQUALITY OF MIXED PARTIAL DERIVATIVES
y(x2 −y 2 )
Exercise 2. f = x2 +y 2 .
2x(x2 + y 2 ) − (2x)(x2 − y 2 ) 4x3 y
f1 = (y) 2 2 2
= 2
(x + y ) (x + y 2 )2
((x2 − y 2 ) − 2y 2 )(x2 + y 2 ) − 2y 2 (x2 − y 2 ) x4 − 4x2 y 2 − y 4
f2 = 2 2 2
=
(x + y ) (x2 + y 2 )2
(x2 + y 2 )2 − 2(x2 + y 2 )(2y)y x4 − 4x2 y 2 − y 4
D2,1 f = 4x3 =
(x2 + y 2 )4 (x2 + y 2 )2
(4x3 − 8xy 2 )(x2 + y 2 )2 − 2(x2 + y 2 )(2x)(x4 − 4x2 y 2 − y 4 )
D1,2 f = =
(x2 + y 2 )4
4x(x2 − 2y 2 )(x2 + y 2 ) − 4x(x4 − 4x2 y 2 − y 4 ) 4xy 2 (3x2 − y 2 )
= =
(x2 + y 2 )3 (x2 + y 2 )3
From the above results, clearly,
lim f1 = 0, lim f1 = 0
x→0 y→0
So that f1 (0, 0) = 0 , while
lim f2 = −1 lim f2 = 1
x→0 y→0
so f2 (0, 0) undefined.
lim f12 = 0 lim f12 = 0
x→0 y→0
so that f12 (0, 0) = 0 , but
4x3 4 x→0
lim f21 = 0 lim f21 = = −−−→ ∞
x→0 y→0 x4 x
92
xy 3
Exercise 3. Given f (x, y) = x3 +y 6 if (x, y) 6= (0, 0), f (0, 0) = 0
a.
!
0 f (x + ha) − f (x) f (ha) − f (0) hax h3 a3y h3 ax a3y
f (0; a) = lim = lim = lim /h = lim =
h→0 h h→0 h h→0 h a3x + h6 a6y
3 h→0 h3 a3
x + h6 a6y
ax a3y a3y
= lim = 2
h→0 a3
x
3
+ h ay 6 ax
a3
So f 0 (0; a) = a2y if ax 6= 0, f 0 (0; a) = 0 if ax = 0
x
b. If x = y 2 , then
xy 3 y5 1 y→0
f (x, y) = 3 6
= 6 = −−−→ ∞ not 0
x +y 2y 2y
So f (x, y) is not continuous at (0, 0).
R √xy −t2
Exercise 4. f (x, y) = 0 e dt for x > 0; y > 0.
√
Let u = u(x, y) = xy and then we can use chain rule.
1p 1p
∂x f = ∂u f ∂x u = e−xy y/x ∂y f = e−xy x/y by label symmetry
2 2
x = x(t)
Exercise 5. Given u = f (x, y); and u = F (t).
y = y(t)
F 0 (t) = (∂x u)x0 + (∂y u)y 0 = x0 ux + y 0 uy
F 00 (t) = x00 ux + x0 (x0 uxx + y 0 uyx ) + y 00 uy + y 0 (x0 uxy + y 0 uyy ) = x00 ux + y 00 uy + x02 uxx + y 02 uyy + (x0 y 0 )(uyx + uxy )
x00 ux + x00 (x0 uxx + y 0 uyx ) + y 000 uy + y 00 (x0 uxy + y 0 uyy )+
F 000 (t) =
+ 2x0 x00 uxx + 2y 0 y 00 uyy + x02 (x0 uxxx + y 0 uyxx ) + y 02 (x0 uxyy + y 0 uyyy )+
+ (x00 y 0 + x0 y 00 )(uyx + uxy ) + (x0 y 0 )(x0 uxyx + y 0 uyyx + x0 uxxy + y 0 uyxy ) =
x00 ux + y 000 uy + x03 uxxx + y 03 uyyy + 3x00 x0 uxx + 3y 00 y 0 uyy +
=
+ 2x00 y 0 uyx + 2y 00 x0 uxy + x0 y 00 uyx + x00 y 0 uxy +
+ x02 y 0 uyxx + y 02 x0 uxyy + x02 y 0 uxyx + x0 y 02 uyxy + x0 y 02 uyyx + x02 y 0 uxxy
x=u+v
Exercise 6. Given f (x, y) into g(u, v), and
y = uv 2
∂f ∂2f ∂2f ∂2f ∂2f
= 2 = 2
= = =1
∂y ∂ x ∂y ∂x∂y ∂y∂x
So
∂u y = v 2 ∂u g = ∂x f ∂u x + ∂y f ∂u y = ∂x f + v 2 ∂y f
∂u x = 1 ∂v g = ∂x f ∂v x + ∂y f ∂v y = ∂x f (1) + 2vu∂y f = ∂x f + 2vu∂y f
2
∂ g
= ∂v (∂x f ) + 2v∂y f + v 2 ∂v (∂y f ) =
∂v∂u
2 2
= (∂xx f + 2vu∂yx f ) + 2v(∂y f ) + v 2 (∂xy
2 2
f + 2vu∂yy f) =
2 2
= (∂xx f + 2vu∂yx f + v 2 ∂xy
2 2
f + 2vu∂yy f + 2v(∂y f )
So for u = 1, v = 1
∂2g
= 1 + 2 + 1(1) + 2(1)(1)(1) + 2(1)(1) = 8
∂v∂u
x = uv
Exercise 7. Given 1
y = (u2 − v 2 )
2
93
(1) Assume equality of mixed partials.
∂g
∂u x = v ∂u y = u = v∂x f + u∂y f
∂u
∂v x = u ∂v y = −v ∂g
= u∂x f − v∂y f
∂v
∂u ∂v g = ∂x f + u∂u (∂x f ) + −v∂u (∂y f ) =
2 2 2 2
= ∂x f + u(v∂xx f + u∂yx f ) + −v(v∂xy f + u∂yy f) =
2 2 2
= ∂x f + uv∂xx f + 2y∂xy f − x∂yy f
(2) Given k∇f (x, y)k2 = (∂x f )2 + (∂y f )2 = 2
2 2
∂g ∂g
a + −b = u2 + v 2 = a(v 2 (∂x f )2 + u2 (∂y f )2 + 2vu∂x f ∂y f ) + −b(u2 (∂x f )2 + v 2 (∂y f )2 − 2uv∂x f ∂y f )
∂u ∂v
=⇒ a = −b since u, v are independent
(∂x f )2 (av 2 + au2 ) + (∂y f )2 (au2 + av 2 ) = a((∂x f )2 + (∂y f )2 )(u2 + v 2 ) = u2 + v 2
=⇒ a = 1/2
Exercise 8. Given that
2F 0 (x)G0 (y)
(F (x) + G(y))2 ez(x,y) = 2F 0 (x)G0 (y); F (x) + G(y) 6= 0 or ez(x,y) =
(F + G)2
so
z(x, y) = ln (2F 0 G0 /(F + G)2 ) = ln F 0 + ln G0 − 2 ln (F + G)
1 2
∂x z = 0 F 00 − F0
F F +G
2 −2F 0 G0
=⇒ ∂yx z= = −ez(x,y) 6= 0
F +G
Exercise 9.
Exercise 11.
(∇f )i = ∂i ((r × A)j (r × B)j ) = ∂i (jkl xk Al )(jmn xm Bn ) =
= jik Ak jmn xm Bn + jkl xk Al jim Bm =
= ijk Aj kmn xm Bn + ijk Bj klm xl Am
So for f (x, y, z) = (r × A) · (r × B),
∇f (x, yz) = B × (r × A) + A × (r × B)
Exercise 12.
(1)
1 −1 1 2xi −xi
∂i = 2 = 3
r r 2 r r
1 −A · r
A·∇ =
r r3
(2)
−aj xj−ai −3 xi −ai 3aj xj xi
∂i = 3 + −aj xj = 3 +
r3 r r4 r r r5
1 −A · B 3(A · x)(x · B)
B·∇ A·∇ = 3
+
r r r5
Exercise 13.
(x − a)2 + (y − b)2 + (z − c)2 = 1 or x2 + y 2 + z 2 − 2xa − 2by − 2zc + a2 + b2 + c2 = 1
consider pts. of intersection, with x2 + y 2 + z 2 = 1 =⇒ 2xa + 2by + 2zc = a2 + b2 + c2
94
Let f (x, y, z) = x2 + y 2 + z 2 − 1 =⇒ ∇f = 2(x, y, z)
Let g(x, y, z) = (x − a)2 + (y − b)2 + (z − c)2 − 1 =⇒ ∇g = 2(x − a, y − b, z − c)
orthogonality condition: ∇g · ∇f = 0 = x(x − a) + y(y − b) + (z − c)z = 1 − ax − by − cz =
2
a + b2 + c2
=1− =0
2
=⇒ 2 = a2 + b2 + c2
√
2 = a2 + b2 + c2 describes a sphere of radius 2 and center at the origin.
Exercise 14.
p
−2x ± 4x2 − 4(1)(y)
2
z= =
z + 2xz + y = 0 =⇒ 2(1)
p
= −x ± x2 − y
Consider parametrizing the position vector for the surface by the x coordinate:
p
r = (x, f (x), −x ± x2 − y)
1
r0 = (1, f 0 , −1 ± p (2x − y 0 ))
2 x2 − y
Now consider the position vector for points contained in the “cylinder.” Note that the z coordinate does not depend upon x
for this cylinder because it looks the same in each x − y plane for each z coordinate.
r = (x, y, z)
r0 = (1, f 0 , 0)
These two tangent vectors must coincide (since the x coordinate and y coordinate are the same, 1 and f 0 , respectively).
1 4x2 − 4xy 0 + y 02
0 = −1 ± p (2x − y 0 ) or 4 =
2 x2 − y x2 − y
=⇒ 4x2 − 4y = 4x2 − 4xy 0 + y 02 or y 02 − 4xy 0 + 4y = 0
p
0 4x ± 16x2 − 4(1)(4y) p
=⇒ y = = 2x ± 2 x2 − y
2
A solution to this ordinary differential equation is y = x2
9.3 E XERCISES - PARTIAL DIFFERENTIAL EQUATIONS , A FIRST- ORDER PARTIAL DIFFERENTIAL EQUATION WITH
CONSTANT COEFFICIENTS
Exercise 1. 4∂x f + 3∂y f = 0
g(3x − 4y) = f (x, y)
4
f (x, 0) = sin x = g(3x) =⇒ g(3x − 4y) = sin (x − y)
3
Exercise 2. 5∂x f − 2∂y f = 0
g(2x + 5y) = f (x, y)
∂x f (x, 0) = 2 g (2x + 5y)|y=0 = 2g 0 (2x) = ex or g(u) = eu/2 + C
0
5
g(2x + 5y) = e(x+ 2 y) + −1
Exercise 3.
(1) If u(x, y) = f (xy), then consider that xy = const. represent level curves for f (because if f (xy) = f (const.), then,
“obviously,” f (const.) = another constant.
Parametrize r by x
+k −k −y k
r = (x, y) = (x, ); r0 = (1, ) = (1, ) where y =
m x2 x x
−y
(∇u) · r0 = ∂x u + ∂y u = 0 or x∂x u − y∂y u = 0
x
95
2
u(x, x) = x4 ex ∂x
2
u(x, x) = f (xx) = f (x2 ) = (x2 )2 ex =⇒ f (xy) = (xy)2 exy
x
(2) v(x, y) = f y for y 6= 0
x
y = const., then f const. ∇f · r0 = 0 on these level curves with xy = const.; r = (x, y) = (x, xk ); r0 = (1, 1/k) or
(1, xy )
y
=⇒ ∂x v + ∂y v = 0 or x∂x v + y∂y v = 0
x
1 0 x y=1/x 1 1
∂x v(x, 1/x) = 1/x; ∂x v = f −−−−→ xf 0 (x2 ) = or f 0 (x2 ) = 2 or f (x) = ln x + C
y y x x
x x
f = ln + C = v(x, y)
y y
x
Since v(1, 1) = 2 =⇒ v(x, y) = ln + 2
y
∂ 2 g(x,y)
Exercise 4. ∂x∂y =0
2
∂y g(x, y) = ψ2 (y) for ∂xy g = 0; g(x, y) = φ2 (y) + φ1 (x) for φ02 (y) = ψ2 (y)
Exercise 5. a = 1, b = −2, c = −3
2 2 2
x = Au + Bv
Consider the general problem: a∂xx f + b∂xy f + c∂yy f =0 , g(u, v) = f (Au + Bv, Cu + Dv)
y = Cu + Dv
∂2g
∂u∂v = 0 (assume equality of mixed partials)
2
∂v g = B∂x f + D∂y f ∂uv g = B(A∂xx f + C∂yx f ) + D(A∂xy f + C∂yy f ) = 0 =
∂u g = A∂x f + C∂y f 2 2 2
= AB∂xx f + (BC + DA)∂xy f + DC∂yy f =0
AB = a
=⇒ BC + DA = b
DC = c
2
∂uv g = 0 =⇒ ∂v g = h(v) or g = H(v) + l(u)
Cx − Ay Dx − By
g(u, v) = H1 (v) + l1 (u) = H1 + l1
BC − AD AD − BC
= H(Cx − Ay) + l(Dx − By)
x+y
Exercise 6. u(x, y) = xyf xy
x+y 0 1 1 −1 x+y −1 0 1 1
∂x u(x, y) = yf + xyf + = yf + yf +
xy y x x2 xy x y x
x2 ∂x u + −y 2 ∂y u = G(x, y)u =⇒
x+y x+y −1
∂y u(x, y) = xf + xyf 0
xy xy y2
=⇒ x2 ∂x − y 2 ∂y u = x2 yf − xyf 0 − y 2 xf + xyf 0 = (x − y)u G=x−y
s
x=e
Exercise 7. f (x, y) =⇒ g(s, t) g(x, t) = f (es , et )
y = et
x2 ∂xx
2
f + y 2 ∂yy
2
f + x∂x f + y∂y f = 0
∂s g = x∂x f ∂t g = y∂y f
2 2 2
∂ss g = x∂x f + x(x∂xx f) ∂tt g = y∂y f + y 2 ∂yy
2
f
2 2
∂ss g + ∂tt g = x∂x f + x2 ∂xx
2
f + y∂y f + y 2 ∂yy
2
f =0
Exercise 8. f (tx) = tp f (x) ∀ t > 0, ∀ x ∈ S s.t. tx ∈ S For fixed x, define g(t) = f (tx);
p
g(t) = f (tx) = t f (x)
=⇒ g 0 (1) = pf (x) = f 0 (x) = (∇f ) · x (by definition of total derivative )
g 0 (t) = ptp−1 f (x)
Exercise 9. Given g(t) = f (tx) − tp f (x), note that we want g(t) = 0.
96
d
g 0 (t) =
f (tx) − ptp−1 f (x)
dt
It is very useful to recall the total derivative definition.
d f ((t + ∆t)x) − f (tx) f (tx + ∆tx) − f (tx)
f (tx) = lim = lim = ((∇f )(tx)) · x
dt ∆t→0 ∆t ∆t→0 ∆t
Use the fact that we’re given: x · (∇f )(x) = pf (x), so that tx · ∇f (tx) = pf (tx)
p g0 p
0
g (t) = x · (∇f )(tx) − pt p−1
f (x) = (f (tx) − tp f (x)) = =
t g t
p =⇒ ln g = p ln t + C
= g(t)
t g = Ktp
Now g(1) = 0 (by plugging into the given g(t) = f (tx) − tp f (x)). But g(1) = 0 if K = 0
=⇒ g = 0 ∀ t
Exercise 10.
g 0 (1) = pf = x∂x f + y∂y f
g 00 (1) = p(x∂x f + y∂y f ) = x∂x (x∂x f + y∂y f ) + y∂y (x∂x f + y∂y f ) =
2 2 2 2
= x(∂x f + x∂xx f + y∂xy f ) + y(x∂yx f + ∂y f + y∂yy f ) = x∂x f + x2 ∂xx
2 2
f + 2xy∂xy f + y∂y f + y 2 ∂yy
2
f
=⇒ x2 ∂xx
2 2
f + 2xy∂xy f + y 2 ∂yy
2
f + (pf ) = p2 f or x2 ∂xx
2 2
f + 2xy∂xy f + y 2 ∂yy
2
f = p(p − 1)f
9.5 E XERCISES - T HE ONE - DIMENSIONAL WAVE - EQUATION
Exercise 4.
2 1 −x x x x 2 −y 2 2y x −y x 2 −y 2
∂xx f= ∂r g + ∂ r g + ∂ g + ∂ g + ∂ θ g + ∂ g + ∂ g
r r2 r r r rr r2 θr r3 r r2 r rθ r2 θθ
2 1 −y y
y y 2
x 2 −2x y x y 2
x 2
∂yy f = ∂r g + ∂r g + ∂ g + ∂ g + ∂θ g + ∂ g + ∂ g
r r2 r r r rr r2 θr r3 r r2 r rθ r2 θθ
2 2 2 ∂r g 2 1 2 1 2 1 2
∂xx f + ∂yy f = ∂r g − + ∂rr g + 2 ∂θθ g = ∂r g + ∂rr g + 2 ∂θθ g=
r r r r r
1 1 2
= ∂r (r∂r g) + 2 ∂θθ g
r r
Exercise 5. We want for
x = ρ cos θ sin φ
y = ρ sin θ sin φ f (x, y, z) → F (ρ, θ, φ)
z = ρ cos φ
x = r cos θ
But first, consider so that f (x, y, z) → g(r, θ, z)
y = r sin θ
(1)
1 1 2 ∂r g 1 2
∇2 f = ∂r (r∂r g) + 2 ∂θθ 2
g + ∂zz g= 2
+ ∂rr g + 2 ∂θθ 2
g + ∂zz g
r r r r
z = ρ cos φ
(2) , so
r = ρ sin φ
1 2 1 2
2
∂θθ g = 2 2 ∂θθ g
r ρ sin φ
Note that, except for a change in notation, this transformation is the same as that used in (a).
2 2 1 1 2
∂zz g + ∂rr g = ∂ρ (∂∂ρ g) + 2 ∂φφ g
ρ ρ
1 1 ρ cos φ 1 cos φ
∂r g = ρ sin φ∂ρ g + 2
∂φ g = ∂ρ g + 2 ∂φ g
r ρ sin φ ρ ρ ρ sin φ
2 1 2 cos φ 1
∇2 f = ∂ρρ2
F + ∂ρ F + 2 ∂φφ F+ 2 ∂ρ F + 2 ∂2 g
ρ ρ ρ sin φ ρ sin φ θθ
97
9.8 E XERCISES - D ERIVATIVES OF FUNCTIONS DEFINED IMPLICITLY, W ORKED EXAMPLES
Exercise 1.
x + y = uv x = X(u, v) xu + yu = v xv + yv = u
∂u : ∂v :
xy = u − v y = Y (u, v) xu y + xyu = 1 xv y + xy + v = −1
1 1 xu v 1 x −1 v xv − 1 1
= =
y x yu 1 x − y −y 1 1 −vy + 1 x−y
1 1 xv u 1 x −1 u ux + 1 1
= =
y x yv −1 x − y −y 1 −1 −uy − 1 x − y
1 1
xu = (xv − 1) xv = (ux + 1)
x−y x−y
−vy + 1 1
yu = yv = (−uy − 1)
x−y x−y
Exercise 5. Given
1 x
F (u, v) = 0 ∂x u = y ∂x v = √ (2x) = √
2 x +z2 2 x + z2
2
u = u(x, y, z) = xy ∂y u = x ∂y v = 0
z
p
2
v = v(x, y, z) = x + z 2 ∂z u = 0
∂z v = √
x + z2
2
x
∂x f = ∂x u∂u F + ∂x v∂v F = y∂u F + ∂v F
v
∂y f = ∂y u∂u F + ∂y v∂v F = x∂u F
z
∂z f = ∂z u∂u F + ∂z v∂v F = ∂v F
v
Since F = f = 0, ∇f · R0 = 0, so ∇f is a normal vector to this surface.
We’re given
√ D1 F (1, 2) = 1
x = 1, y = 1, z = 3
D2 F (1, 2) = 2
so then
x z
∇f = (y∂u F + ∂v F, x∂u F, ∂v F ) =
v √ v
1 3 √
= (1 + 2, 1(1), 2) = (2, 1, 3)
2 2
√ r !
(2, 1, 3) 1 1 1 3
=⇒ √ = √ , √ ,
2 2 2 2 2 2 2
Exercise 6.
x=y=1
x2 − y cos (uv) + z 2 = 0 x = x(u, v)
∂u x u = π/2
x2 + y 2 − sin (uv) + 2z 2 = 2 y = y(u, v) we want at
∂v x v=0
xy − sin u cos v + z = 0 z = z(u, v)
z=0
2xxu − yu cos (uv) + y sin (uv)v + 2zzu = 0
∂u : 2xxu + 2yyu − cos (uv)(v) + 4zzu = 0 or
xu y + xyu − cos u cos v + zu = 0
(2x, 2x, y)xu + (− cos (uv), 2y, x)yu + (2z, 4z, 1)zu = (−y sin (uv)v, v cos (uv), cos u cos v)
−y sin (uv)v v cos (uv) cos u cos v
− cos (uv) 2y x
2z 4z 1
xu =
2x 2x y
− cos (uv) 2y x
2z 4z 1
Note that − y sin (uv)v = v cos (uv) = cos u cos v = 0
=⇒ xu = 0
98
2xxv − yv cos (uv) + y sin (uv)u + 2zzv = 0
∂v : 2xxv + 2yyv − cos (uv)u + 4zzv = 0 or
xv y + xyv + sin u sin v + zv = 0
(2x, 2x, y)xv + (− cos (uv), 2y, x)yv + (2z, 4z, 1)zv = (−y sin (uv)u, cos (uv)u, − sin u sin v)
−yu sin (uv) u cos (uv) − sin u sin v 0 π/2 0
− cos (uv) 2y x −1 2 1
2z 4z 1 0 0 1
xv = =
2x 2x y 2 2 1
− cos (uv) 2y x −1 2 1
2z 4z 1 0 0 1
= π/12
9.13 E XERCISES - M AXIMA , MINIMA , AND SADDLE POINTS . S ECOND - ORDER TAYLOR FORMULA FOR SCALAR
FIELDS . T HE NATURE OF A STATIONARY POINT DETERMINED BY THE EIGENVALUES OF THE H ESSIAN MATRIX .
S ECOND - DERIVATIVE TEST FOR EXTREMA OF FUNCTIONS OF TWO VARIABLES .
Exercise 1. z = x2 + (y − 1)2
f (x, y) = x2 + (y − 1)2
f ≥ 0 and f = 0 when (x, y) = (0, 1)
∇f = (2x, 2(y − 1)) = 0 where (x, y) = (0, 1)
(0, 1) is an abs. min.
Exercise 2. z = x2 − (y − 1)2
For (x, y) = (t, 1), f (t, 1) = t2 ≥ 0
∇f = (2x, −2(y − 1)) = 0 when (x, y) = (0, 1)
For (x, y) = (0, 1 + δ), f (0, 1 + δ) = −δ 2 < 0
So (0, 1) is a saddle pt.
Exercise 3. z = 1 + x2 − y 2
For (x, y) = (0, u), z = 1 + −u2 ≤ 1
∇f = (2x, −2y) = 0 when (x, y) = (0, 0)
For (x, y) = (t, 0), z = 1 + t2 ≥ 1
So (0, 0) is a saddel pt.
Exercise 4. z = (x − y + 1)2
∇f = (2(x − y + 1), 2(x − y + 1)(−1)) = 0 when y = x + 1
f ≥ 0 ∀ (x, y), so (x, x + 1) is an abs. min. since f (x, x + 1) = 0
Exercise 5. z = 2x2 − xy − 3y 2 − 3x + 7y
y = 4x − 2 x = 3/7
∇f = (4x − y − 2, −x + 2y + 1) = 0 where x − 1 so
y= y = −2/7
2
4 −1
H= =⇒ (λ − 4)(λ − 2) − 1 = (λ − 7)(λ + 1)
−1 2
So we have a saddle point at (3/7, −2/7).
Exercise 6. For z = x2 − xy + y 2 − 2x + y,
∇f = (2x − y − 2, −x + 2y + 1) = 0
so the critical point is at (x, y) = (1, 0).
The Hessian matrix is
2 −1
H=
−1 2
So λ = 1, +3 are the eigenvalues. (1, 0) is a relative minimum.
Exercise 7. For z = x3 − 3xy 2 + y 3 ,
∇f = (3x2 − 3y 2 , −6xy + 3y 2 ) = 3(x2 − y 2 , −2xy + y 2 ) = 0
Then because y 2 − 2xy = 0, (x, y) = (0, 0).
99
The Hessian matrix is
6x −6y
H=
−6y −6x + 6y
For (x, y) = (0, 0), by theorem, the Hessian matrix doesn’t give a definite conclusion. Then resort to the definitions of saddle
points, relative minima, relative maxima.
Now z = x3 − 3xy 2 + y 3 = y 2 (y − 3x) + x3 . z(0, 0) = 0.
Consider (x, y) = (δ, ). Then
z = 2 ( − 3δ) + δ 3
So for y = 3x,
∀ E > 0, ∃δ > 0 s.t.
if 0 < x < δ, then 0 < z < 2δ = E(δ),
if −δ < x < 0, then −E(δ) = −2δ 3 < z < 0
So in the neighborhood of (0, 0), there exist pts. above and below z = 0. By definition, (0, 0) is a saddle point.
Exercise 8. For z = x2 y 3 (6 − x − y),
∇f = (12xy 3 − 3x2 y 3 − 2xy 4 , 18x2 y 2 − 3x3 y 2 − 4x2 y 3 ) = 0
With
12xy 3 − 3x2 y 3 − 2xy 4 = xy 3 (12 − 3x − 2y) = 0
x2 y 2 (18 − 3x − 4y) = 0
So (x, y) = (2, 3) or (x, 0), (0, y) are the critical points.
The Hessian matrix is
12y 3 − 6xy 3 − 2y 4 36xy 2 − 9x2 y 2 − 8xy 3
H=
36xy 2 − 9x2 y 2 − 8xy 3 36x2 y − 6x3 y − 12x2 y 2
Now
A = D1,1 f = 2y 3 (6 − 3x − y)
B = D1,2 f = xy 2 (36 − 9x − 8y)
C = D2,2 f = 6x2 y(6 − x − 2y)
For (2, 3), ∆ < 0. (2, 3) is a saddle point.
Looking at the Hessian matrix, the definitions must be used to determine if the critical points are saddle points, relative
maxima, or relative minima.
Consider (x, 0), z(x, 0) = 0.
Consider |y| < δ2 .
z = y 3 x2 (6 − x − y)
Choose δ2 , s.t. δ2 < |6 − x| (since δ2 is arbitrarily small, we can make this choice).
Then for fixed x, either 6 − x < 0, or 6 − x > 0.
But for |y| < δ2 , y > 0, or y < 0, either z < 0, z > 0, since y or −y allowed, for |y| < δ2 .
|z| = |y 3 ||x2 ||6 − x − y| < δ23 2|6 − x|x2 = E(x, δ2 )
∀ E > 0, ∃δ2 > 0 s.t. for |y| < δ2 , |z| < E(x, δ2 ) and in this neighborhood, ∃ (x, y) s.t. z < 0 and (x, y) s.t. z > 0.
(x, 0) are saddle points.
(0, y), z(0, y) = 0.
For |x| < δ1 ,
z = x2 y 3 (6 − y − x).
Consider δ1 < |6 − y|.
For y < 0, y > 6, z < 0 for infinitesimal neighborhood about z (with δ1 < |6 − y|).
For (0, y), y < 0, y > 6, z(0, y) a relative minimum.
Likewise, for 0 < y < 6, z > 0 for infinitesimal neighborhood about z, (with δ1 < |6 − y|), so z(0, y) a relative maximum.
For (0, 6), z(x, 6) = 216x2 (−x) = −216x3 . ∀ E > 0, ∃ δ1 > 0 s.t. |z| < E when |x| < δ1 ,
For |x| < δ1 , both x, −x fulfill the condition, so that
∃ pts. (x, 6), (−x, 6) in this neighborhood such that z < 0, z > 0, respectively.
(0, 6) a saddle point.
Exercise 9. z = x3 + y 3 − 3xy
100
2 2
x2 = y
∇f = (3x − 3y, 3y − 3x) = 0 =⇒ =⇒ (0, 0), (1, 1)
y2 = x
6x −3
|Df | = = 36xy − 9
−376y
(1, 1) minimum since Df (1, 1) = 27 and D1,1 f (1, 1) = 6 > 0
(0, 0) saddle pt. since Df (0, 0) = −9 < 0
Exercise 10. z = sin x cosh y
− sin x cosh y cos x sinh y
∇f = (cos x cosh y, sin x sinh y) |Df | = = − sin2 x cosh2 y + − cos2 x sinh2 y
cos x sinh y sin x cosh y
2j − 1
∇f = 0 =⇒ (x, y) = π, 0
2
2j − 1 2j − 1
Df π, 0 = −1 =⇒ π, 0 is a saddle pt.
2 2
Exercise 11. z = e(2x+3y) (8x2 − 6xy + 3y 2 ) = f eg . It helps alot to make these notation substitutions.
fx = 16x − 6y
∇z = (fx eg + 2f eg , fy eg + 3f eg )
fy = −6x + 6y
∇z (fx + 2f )eg = 0 −1 −1
−−→ =⇒ (0, 0), ( , )
(fy + 3f )eg = 0 4 2
16eg + 4fx eg + 4f eg −6eg + 2fy eg + 3fx eg + 6f eg
Dij z =
−6eg + 3fx eg + 2fy eg + 6f eg 6eg + 6fy eg + 9f eg
16 −6
Dij z(0, 0) = = 96 − 36 = 60 (0, 0) is a minimum
−6 6
−1 −1 16 −9 −1 −1
Dij z( , ) = e−4 −3 = e
−4
(−24 − 54) < 0 ( , ) is a saddle pt.
4 2 −6 2 4 2
2
+xy+y 2 )
Exercise 12. z = (5x+7y−25)e−(x = f e−g . Using these shorthand, substitution notation helps with the calculation.
∇z = (5e−g + f e−g (−2x − y), 7e−g + f e−g (−x − y))
5 + (−2x − y)f = 0
−1 −3
∇z = 0 =⇒ =⇒ (x, y) = (1, 3), ,
7 + (−x − 2y)f = 0 26 26
Dij z =
2
−g 5(−2 − y)2 + f (−2x − y) + f (−2) 7(−2x − y) + 5(−x − 2y) + f (−x − 2y)(−2x − y) − f
e
7(−2x − y) + 5(−x − 2y) + f (−x − 2y)(−2x − y) − f 7(−x − 2y)2 + (−x − 2y)2 f + −2f
−27 −36
Dij z(1, 3) = e−2g > 0 =⇒ (1, 3) is a maximum
−36 −51
25
+ 52 35 + 26
−1 −3 26 26 > 0 =⇒ ( −1 , −3 ) is a minimum.
Dij z( , ) = 35
49
26 26 26 + 26 26 + 52 26 26
Exercise 13. z = sin x sin y sin (x + y), 0 ≤ x ≤ π, 0 ≤ y ≤ π
Exercise 21. Method of least squares. Given n distinct numbers x1 , . . . , xn and n further numbers y1 , . . . , yn , and f (x) =
ax + b fitting form,
n
X n
X
E(a, b) = (f (xi ) − yi )2 = (axi + b − yi )2
i=1 i=1
n
X X X X
∂a E = 2(axi + b − yi )xi = 2 a x2i + b xi − yi xi = 0
i=1
∇E = 0 =⇒ n X
X X
∂b E = 2(axi + b − yi ) = 2 a xi + nb + − yi = 0
i=1
101
X
X2 = x2i
1X X2
P
X= xi nX a yi xi
=⇒ =
n nX n b nY
1 X
Y = yi
n
yi xi − n2 XY
P P
a 1 n −nX yi xi n P 1
=⇒ = =
b nX 2 − n2 X
2 −nX X2 nY −nX yi xi + nX 2 Y nX 2 − n2 X
2
ui = xi − X
2 X 2 2 2
Suppose u2i = x2i − 2xi X + X =⇒ u2i = X 2 − 2nX + nX = X 2 − nX
X X X
yi ui = yi (xi − X) = yi xi − XY n
X X
=⇒ a = yi ui / u2i
Then use anX + nb − nY = 0 or b = Y − aX toPget b.
n Pn
Exercise 22. f (x, y) = ax + by + c. E(a, b, c) = i=1 (f (xi , yi ) − zi )2 = i=1 (axi + byi + c − zi )2 . (xi , yi ) are n given
distinct pts. z1 , . . . , zn are n given real numbers.
X X
X X X
∂a E = 0 = 2 (axi + byi + c − zi )xi =⇒ x2i + b
a xi yi + c xi − zi xi = 0
X X X X X
∂b E = 0 = 2 (axi + byi + c − zi )yi =⇒ a xi yi + b yi2 + c yi − zi yi = 0
X X X X
∂c E = 0 = 2 (axi + byi + c − zi ) =⇒ a xi + b yi + nc − zi = 0 or c = Z − aX − bY
Then rewrite the above equations substituting the expression for c.
X X X
aX 2 + b xi yi + cnX = zi xi = aX 2 + b xi yi + nX(Z − aX − bY ) =
2 X
= a(X 2 − nX ) + b( xi yi − nXY ) + nXZ
X X X 2
a xi yi + bY 2 + cnY = zi yi = a( xi yi − nXY ) + b(Y 2 − nY ) + nY Z
Then
" 2
# P
X 2 − nX
P P
xi yi − nXY a z x − nXZ uz
2 = P i i = P i i
Y 2 − nY b zi yi − nY Z vi zi
P
xi yi − nXY
2 X
X 2 − nX = u2i
Note that for ui = xi − X, vi = yi − Y , we already showed in the previous exercise, Exercise 21, that 2 X
Y 2 − nY = vi2
X X X X
u i vi = (xi − X)(yi − Y ) = xi yi − nXY − nXY + XY n = xi yi − nXY
P 2 P
u Pui 2vi and use Cramer’s rule to obtain
So let ∆ = P i
ui v i vi
P P
1 ui zi
Pui 2vi
a= P
∆ vi zi vi
P P
1 vi zi ui v i
b= P P 2
∆ ui zi ui
c = Z − aX − bY
Exercise 23. z1 , . . . , zn are n distinct pts. in m-space.
102
Pn Pn Pm
Let x ∈ Rm , let f (x) = k=1 kx − zk k2 = k=1 j=1 (xj − (zk )j )2
n
X
Now ∂i f = 2(xi − (zk )i )
k=1
n
X n
X
Conditions we want: ∇f = 0 =⇒ (xi − (zk )i ) = 0 =⇒ nxi − (zk )i = 0
k=1 k=1
n n
1 X 1X
=⇒ xi = (zk )i or x = zk
n n
k=1 k=1
n
X n
X
∂ij f = ∂i 2(xj − (zk )j ) = 2(1)δij = 2nδij
k=1 k=1
1
Pn
H is diagonalized and λj > 0, ∀ j = 1, ,̇m. By Thm., a = n k=1 zk , (the centroid) is a minimum.
Exercise 25. f (x, y, z) = x4 + y 4 + z 4 − 4xyz.
x3 = yz
∇f = (4x3 − 4yz, 4y 3 + −4xz, 4z 3 − 4xy) = 0 = (x3 − yz, y 3 − xz, z 3 − xy) =⇒ y 3 = xz thus ∇f (1, 1, 1) = 0
3
z = xy
12x2
2
−4z −4y 3x −z −y 3 −1 −1
H = −4z 12y 2 −4x = 4 −z 3y 2 −x H(a) = 4 −1 3 −1
−4y −4x 12z 2 −y −x 3z 2 −1 −1 3
λ − 12 4 4
4 = (λ − 16)2 (λ − 4) = 0
|λI − H| = 4 λ − 12
4 4 λ − 12
=⇒ λ = 16, 4
a = (1, 1, 1) is a minimum, by theorem.
9.15 E XERCISES - E XTREMA WITH CONSTRAINTS . L AGRANGE ’ S MULTIPLIERS
Exercise 1. Given f (x, y) = z = xy and g(x, y) = 0 = x + y − 1
∇f = (y, x)
=⇒ ∇f = λ∇g = (y, x) = λ(1, 1) = (1 − x, x)
∇g = (1, 1)
1
x= 1
=⇒ 1 − x = x so that 2 z=
1 4
y=
2
Exercise 2.
p (x, y)
f (x, y) = r = x2 + y 2 ∇f == λ(10x + 6y, 10y + 6x)
2 2
r
g(x, y) = 5x + 6xy + 5y − 8 ∇g = (10x + 6y, 10y + 6x)
(x, y) x 1 y 1
∇f = = λ(10x + 6y, 10y + 6x) =⇒ = or x2 = y 2
r r 10x + 6y r 10y + 6x
1 √
g(x, ±x) = 5x2 ± 6x2 + 5x2 − 8 = 0 =⇒ 10x2 ± 6x2 = 8 or x = ± √ , ± 2
2
√ √ √ √
r maximum , 2, when ( 2, ± 2), (− 2, ± 2)
1 1 1 1
r minimum , 1, when ( √ , ± √ ), ( √ , ± √ )
2 2 2 2
Exercise 3. a, b > 0
(1)
x y
f =z= +
a b g(x, y) = x2 + y 2 = 1
1 1 ∇g = (2x, 2y)
∇f = ,
a b
103
1 ±b
= λ2x y x= √
a2
a a2 + b2
∇f = λ∇g =⇒ a or = so then x2 1+ 2 = 1 or
1 x b b ±a
= λ2y y=√
b a2 + b2
√ √
b2 + a2 b2 + a2
z= ,−
ab ab
Geometrically, consider lines of bz − ab x = y inside a circular region of x2 + y 2 = 1.
(2)
x y
g(x, y) = + −1
f = z = x2 + y 2
a b 1 1 a y
1 1 ∇f = λ∇g =⇒ 2(x, y) = λ , or =
∇f = 2(x, y) ∇g = , a b b x
a b
ab2
x=
ab2 ba2 a2 b2
a2 + b2
Plug back into g(x, y): =⇒ minimum at , z=
ba2 b2 + a2 a2 + b2 a2 + b2
y= 2
a + b2
bx
Geometrically, consider points on a line defined by g(x, y), y = b − a . Then f defines circles of increasing radius.
Obviously, we can make the radius for f , z, as large as we want.
Exercise 4.
Exercise 5.
f (x, y, z) = x − 2y + 2z g = x2 + y 2 + z 2 − 1 = 0 ∇f =λ∇g 1 −1 1
−−−−−−→ = =
∇f = (1, −2, 2) ∇g = (2x, 2y, 2z) 2x y z
2x = z ±2
=⇒ z =
z = −y 3
1 −2 2
f , , =3
3 3 3
−1 2 −2
f , , = −3
3 3 3
Exercise 6.
p
f= x2 + y 2 + z 2 g = z 2 − xy − 1
(x, y, z) ∇g = (−y, −x, 2z)
∇f =
4
x2 = y 2
∇f =λ∇gx y 1
−−−−−−→ = = =⇒ −x so x = y = 0
−y −x 2 y=
2
=⇒ z = 1, −1 or (0, 0, ±1) are the points where the distance is minimized.
Exercise 7.
p (x − 1)
f (x, y) = (x − 1)2 + y 2 = 4λ
g(x, y) = 4x − y 2
p
∇f =λ∇g (x − 1)2 + y 2
(x − 1, y) −−−−−−→
∇f = p ∇g = (4, −y) y
(x − 1)2 + y 2 p = −2yλ
(x − 1)2 + y 2
x > 0, but if y 6= 0, the equations imply x = −1.
=⇒ (x, y) = (1, 0) is the point of shortest distance on the parabola to (1, 0)
Exercise 8. Given the constraining surfaces
x2 − xy + y 2 − z 2 = 1
or xy + z 2 = g = 0
x2 + y 2 = 1
104
(x,y,z)
p
Then we want to minimize f = x2 + y 2 + z 2 . ∇f = f
x
= λy
f
∇f =λ∇g y
∇g1 = (y, x, 2z) −−−−−−→ = λx
f
z
= λ2z
f
y x x
= y=
1 f 2f 2 Contradiction.
6 0, then
Suppose z = = λ =⇒ x 1 or y
2f = y x=
f 2f 2
Then z = 0.
x y
Suppose x, y 6= 0 =⇒ = or x2 = y 2
y x
±1
=⇒ 2x2 = 1 or x = √ but z 2 = −xy = 0 Contradiction.
2
Then x = 0 or y = 0, so that λ = 0
(0, ±1, 0), (±1, 0, 0)
Exercise 9.
f (x, y, z) = xa y b z c
g =x+y+z−1 ∇f =λ∇g af bf cf
−−−−−−→ = =
a b c ∇g = (1, 1, 1)
∇f = f , , x y z
x y z
If x, y, z 6= 0 then
bx
y= aa bb cc
a =⇒ x + bx + cx = 1 so that the maximum occurs at (x, y, z) = 1
(a, b, c) and f =
cx a a a+b+c (a + b + c)a+b+c
z=
a
2
x2 2
Exercise 10. Consider the ellipsoid + yb2 + zc2 − 1 = g1 . Then ellipsoid will have the same normal at a point on the ellipsoid
a2
as the tangent plane through the same point. Thus, we want
x y z
∇g1 = 2 2 , 2 , 2
a b c
to be a normal that defines a plane through (x, y, z), that’s on the ellipsoid, in the uvw plane.
xu yv zw
=⇒ 2 + 2 + 2 = 1
a b c
The volume of the tetrahedron is
Z a2 /x Z by2 (1− xu2 ) Z cz2 (1− xu2 − yv2 )
(abc)2 (abc)2
a a b −1 −1 −1
V = du dv dv = ∇V = , ,
0 0 0 6xyz 6 x2 yz xy 2 z xyz 2
(abc)2
−1 x a
2
= λ2 2 x= √
6 x yz a 2
bx 3
(abc) 2
−1
y y2 = x2
+ y2
+ z2
=1 b
∇V =λ∇g1 a 2 2 2
−−−−−−−→ = λ2 2 so −a−−−b−−−c−−→ y = √
6 xy 2 z b c 2 3
z2 = x c
(abc)2
−1 z a z=√
= λ2 2 3
6 xyz 2 c
√
3
V = abc
2
Exercise 12. Consider the conic section as a quadratic form.
2 2 2 A B
Ax + 2Bxy + Cy = 1 where A > 0 and B < AC =⇒ T =
B C
Note that detT = AC − B 2 , so normalize T by detT so to obtain only pure rotation, no “amplification.”
p
(A + C) ± (A − C)2 + 4B 2
1 λ−A −B
T →T = = 0 =⇒ λ± =
AC − B 2 −B λ−C 2(AC − B 2 )
105
We don’t need to find the eigenvectors. By theorem, we can find a C s.t. Y = XC and Y are the coordinates in which T is
diagonalized.
u2 v2
λ+ u2 + λ− v 2 = 1 =⇒ + =1
1/λ+ 1/λ−
Immediately we recognize this to be the equation of an ellipse. The T rotation does not amplify distances since detT = 1, and
so distances from the origin to the conic section are preserved. Then we can immediately name the minimum and maximum
distances: p p
2 2 (A + C) + (A − C)2 + 4B 2 (A + C) − (A − C)2 + 4B 2
M , m = 1/ , 1/
2(AC − B 2 ) 2(AC − B 2 )
Exercise 13. Let X = (x, y) be a point on the ellipse.
The line is given by the set {(0,√
4) + s(1, −1)|s ∈ R}.
The normal to the line is (1, 1)/ 2, so connect a point on the ellipse to the line by a perpendicular distance t by the following:
(1, 1)
X = t √ = (0, 4) + s(1, −1)
2
so that
t(1, 1)
√ = (0, 4) + (s, −s) − (x, y) = (s − x, 4 − s − y)
2
t
√ =s−x
2 √ 4−x−y
=⇒ 2t = 4 − x − y so let f (x, y) = t = √
t 2
√ =4−s−y
2
x2
−1 −1
g(x, y) = + y2 − 1 = 0 x
∇f = √ , √ 4
x =⇒ = 2y =⇒ 16y 2 /4 + y 2 = 5y 2 = 1
2 2 2
∇g = , 2y
2
−4 √
Thus, the points that extremize f are (x, y) = √4 , √1
5 5
, √
5
, −15 .
√ √
4− 5 4+ 5
tmin = √ , tmax = √
2 2
10.5 E XERCISES - I NTRODUCTION , PATHS AND LINE INTEGRALS , OTHER NOTATIONS FOR LINE INTEGRALS , BASIC
PROPERTIES OF LINE INTEGRALS
Exercise 1. f = ((x2 − 2xy), (y 2 − 2xy)) from (−1, 1) to (1, 1); y = x2
α(x) = (x, x2 ); α0 = (1, 2x)
Z Z
((x − 2xy), (y − 2xy)) · (1, 2x)dx = (x2 − 2xy + 2xy 2 − 4x2 y)dx =
2 2
1
1 2 3 5 4 1 3 2 4 2 6 4 5 1 4 2 8 −14
−1 (x − 2x + 2x − 4x )dx = x − x + x − x = (1 − (−1)) − (1 − (−1)) = − =
3 4 6 5 −1 3 5 3 5 15
Exercise 2. f = (2a − y, x) along the path described by α = (a(t − sin t), a(1 − cos t)) 0 ≤ t ≤ 2π
f (t) = (2a − a(1 − cos t), a(t − sin t)) = (a + a cos t, a(t − sin t))
α0 (t) = (a(1 − cos t), a sin t)
Z Z 2π Z 2π
f (t) · α0 (t)dt = (a2 (1 − cos2 t) + a2 (t sin t − sin2 t))dt = a2 (1 + t sin t − 1)dt =
0 0
Z 2π Z 2π
2π
= a2 t sin tdt = a2 (−t cos t)|0 − − cos tdt = a2 (−2π) + 0 = −2πa2
0 0
Exercise 3. f (x, y, z) = ((y 2 − z 2 ), 2yz, −x2 ); α(t) = (t, t2 , t3 ), 0≤t≤1
f [α(t)] = ((t4 − t6 ), 2t5 , −t2 ) α0 (t) = (1, 2t, 3t2 )
Z 1 Z 1 1
0 4 6 6 4 1 5 1 7 −2 3 1
f (α(t)) · α (t)dt = (t − t + 4t − 3t )dt = (−2) t + 3t = + =
0 0 5 7 0 5 7 35
Exercise 4. f = (x2 + y 2 , x2 − y 2 ) from (0, 0) to (2, 0) along the curve y = 1 − |1 − x|.
106
( (
1−x if 1 − x > 0 or 1 > x x 1>x
|1 − x| = y=
−(1 − x) if 1 − x < 0 or 1 < x 2−x 1<x
Forx < 1, For x > 1,
α(x) = (x, x) α(x) = (x, (2 − x))
0
α (x) = (1, 1) α0 (x) = (1, −1)
Z Z 1 Z 2 Z 1 Z 2
f · α0 (x)dx = ((x2 + y 2 ) + (x2 − y 2 ))dx + ((x2 + y 2 ) − (x2 − y 2 ))dx = 2x2 dx + 2(2 − x)2 dx =
0 1 0 1
1 2
2 3 1 3 2 −2 2 2 4
(0 − 13 ) = + =
= x +2 (2 − x) (−1) = +
3 0 3 1 3 3 3 3 3
2 2
=⇒ xa + yb = 1. From the ellipse equation, parametrize by
Exercise 5. f = (x + y, x − y) b2 x2 + a2 y 2 = a2 b2
x = a cos θ
θ. =⇒
y = b sin θ
α(θ) = (a cos θ, b sin θ)
α0 (θ) = (−a sin θ, b cos θ)dθ
Z 2π Z 2π
0
f (θ) · α (θ)dθ = ((a cos θ + b sin θ)(−a sin θ)dθ + (a cos θ − b sin θ)b cos θdθ) =
0 0
Z 2π Z 2π
2
= (−a − b ) 2
cos θ sin θdθ + ab (cos2 θ − sin2 θ)dθ =
0 0
2π 2π
sin2 θ
2 2 sin 2θ
= (−a − b ) + ab = 0
2 0 2 0
Exercise 6. Given f = (2xy, x2 + z, y) from x1 = (1, 0, 2) to x2 = (3, 4, 1), use vector calculus and analytic geometry to
form a line with direction vector P = x2 − x1 = (2, 4, −1). Thus, the line is described by x = tP + x1 , where t is the
parameter.
x = 2t + 1
x = tP + x1 =⇒ y = 4t
z = −t + 2
f (t) = (2(2t + 1)(4t), ((2t + 1)2 + (−t + 2)), 4t) = (16t2 + 8t, 4t2 + 3t + 3, 4t)
α0 (t) = (2, 4, −1)
Z 1 Z 1 Z 1
f (t) · α0 (t)dt = (32t2 + 16t + 16t2 + 12t + 12 − 4t)dt = dt(48t2 + 24t + 12) =
0 0 0
1
48 3 24 2
= t + t + 12t = 16 + 12 + 12 = 40
3 2 0
Exercise 7. f (x, y, z) = (x, y, (xz − y)) from (0, 0, 0) to (1, 2, 4) along a line segment.
x1 = (0, 0, 0) x t
P = x2 − x1 = x2 . The line is described by P t + x1 = x = P t, so y = 2t
x2 = (1, 2, 4) z 4t
Z 1 Z 1 1
−3t2 16t3
Z Z
23
f (t)·α0 (t)dt = (t+4t+(4t2 −2t)4)dt = (t+4t+16t2 −8t)dt = (−3t+16t2 )dt = + = 6
0 0 2 2 0
Exercise 8. Given f (x, y, z) = (x, y, (xz − y)) along the path described by α(t) = (t2 + 2t, 4t3 ); 0≤t≤1
α0 (t) = (2t, 2, 12t2 )
f (t) = (t2 , 2t, (4t2 t3 − 2t)) = (t2 , 2t, (4t5 − 2t))
Z 1 1
2 4 4 2 48 8 24t4
Z
5
f (t) · α0 (t)dt = (2t3 + 4t + 12t2 (4t5 − 2t))dt = t + t + t − = 2
0 4 2 8 4 0
Exercise 9. Given C (x − 2xy)dx + (y − 2xy)dy; where C is a path from (−2, 4) to (1, 1) along the parabola y = x2
2 2
R
107
parametrize to x.
1 1
1 3 2 4 2 6 4x5
Z Z
((x2 − 2x(x2 ))dx + (x4 − 2x(x2 ))2xdx) = 2 3 5
(x − 2x + 2x − 4x )dx = 4
x − x + x − =
C −2 3 4 6 5 −2
1 1 1 4 15 63 4 −369
= (1 − (−8)) − (1 − 16) + (1 − 64) − (1 − (−32)) = 3 + − − (33) =
3 2 3 5 2 3 5 10
x dx
= cos θ = −a sin θ
(x+y)dx−(x−y)dy x 2 y 2
= 1. Let a dθ
where C is the circle x2 + y 2 = a2 or
R
Exercise 10. C x2 +y 2 a + a y
= sin θ dy
a = a cos θ
dθ
Z 2π
=⇒ ((a cos θ + a sin θ)(−a sin θ)dθ − (a cos θ − a sin θ)a cos θdθ) /a2 = (−1)(2π) = −2π
0
dx+dy
R
Exercise 11. F |x|+|y|
, where F = A + B + C + D, where
A : y = −x + 1
B : y =x+1
C : y = −x − 1
D : y =x−1
So Z
dx −dx
+ =0
A x + (−x + 1) x + (−x + 1)
Z Z −1
2dx dx
=2 = 2(−1) = −2 Z
−x + x + 1 1 dx + dy
ZB 0
=⇒ =0
dx − dx F |x| + |y|
=0
C −x +x+1
Z Z 1
2dx dx
=2 = 2(1) = 2
D x + (−x + 1) 0 1
Exercise 12.
ydx+zdy +xdz, on the intersection of x+y = 2 or y = 2−x and x2 +y 2 +z 2 =
R
(1) Given that we want to compute C
2(x + y) = 2(2) = 4, then
dx = dx
dy = −dx
dz −4x + 4
2z = −4x + 4 or dz = √ dx
dx 2 4x − 2x2
and
z 2 = 4 − x2 − y 2 = 4 − x2 − (2 − x)2 = −2x2 + 4x
2 Z 0
2(1 − x) x2(1 − x)
Z p p
(2 − x)dx + 4x − 2x (−dx) + x √
2 dx + (2 − x)dx + − 4x − 2x (−dx) + √
2 dx =
0 4x − 2x2 2 − 4x − 2x2
Z 2 Z 2
−(4x − 2x2 ) + 2x − 2x2 p
=2 √ dx = (−4) (x/ 4x − 2x2 )dx
0 4x − 2x2 0
x
Let u = 2 . Then 2du = dx.
√
Z 1 √ p √ Z 1 u √ Z π/2 sin θ
=⇒ (−4) (2u)(2du)/ 2 4u − (2u)2 = −4 2 √ du = −4 2 2 sin θ cos θdθ =
0 0 1−u 0 cos θ
√
1 π − sin 2θ π/2 √
= −8 2 −9 = −2 2π
2 2 4
0
u = sin2 θ
since and
du = 2 sin θ cos θdθ
√ p
1 − u = 1 − sin2 θ = cos θ
108
(2) With x2 + y 2 = 1, z = xy, let x = cos θ, y = sin θ, z = cos θ sin θ, so that
Z Z 2π
ydx + zdy + xdz = (sin θ(− sin θ)dθ + cos θ sin θ cos θdθ + cos θ(− sin2 θ + cos2 θ)dθ =
C 0
Z 2π
= (− sin2 θ + cos2 θ sin θ − sin2 θ cos θ + cos3 θ)dθ =
0
2π
− cos3 θ 1
1 − cos 2θ 1
− sin3 θ + sin θ − sin3 θ = −π
= (−1) +
2 3 3 3 0
10.9 E XERCISES - T HE CONCEPT OF WORK AS A LINE INTEGRAL , L INE INTEGRALS WITH RESPECT TO ARC LENGTH ,
F URTHER APPLICATIONS OF LINE INTEGRALS
Exercise 1. f (x, y, z) = (x, y, (xz − y))
x1 = (0, 0, 0) P = x2 − x1 = x2
x2 = (1, 2, 4)x = tx2 + x1 ; 0 ≤ t ≤ 1
1 Z 1
−3 16
Z Z Z
ds 23
f · ds = f · dt = (t, 2t, t(4t) − 2t) · (1, 2, 4)dt = (t + 4t + 16t2 − 8t)dt = + =
dt 0 0 2 3 6
Exercise 2. f (x, y) = ((x2 − y 2 ), 2xy)
A : r = (a, ta)
B : r = (a(1 − t), a)
C : r = (0, a(1 − t))
D : r = (at, 0)
Z 1 Z 1
A: (a2 − a2 t2 , 2ta2 ) · (0, a)dt = 2ta3 dt = a3
0 0
Z 1 Z 1 1
1 2
((a2 (1 − 2t + t2 ) − a2 ), 2a(1 − t)a) · (−a, 0)dt = −a a2 (−2t + t2 )dt = −a3 −t2 + t3 = a3
B:
0 0 3 0 3
Z 1
C: (−a2 (1 − 2t + t2 ), 0) · (0, −a) = 0
0
Z 1
1
D: (a2 t2 , 0) · (a, 0)dt = a3
0 3
Z
=⇒ f · ds = 2a3
C
Exercise 3. f (x, y) = (cxy, x6 y 2 ) c > 0. (0, 0) to line x = 1 via y = axb ; a > 0, b > 0.
Z Z Z
b 6 2 2b b−1
W = f · ds = (cx(ax ), x a x ) · (1, bax )= (acxb+1 + a3 bx3b+5 )dx =
ac a3 b 3ac a3 b 3ac + a3 b
= + = + =
b + 2 3b + 6 3(b + 2) 3(b + 2) 3(b + 2)
r
3c 3c
=⇒ 2 = 2 =⇒ a =
a 2
a = (0, 0, 0)
Exercise 4. f = (yz, xz, x(y + 1)). b = (1, 1, 1)
c = (−1, 1, −1)
109
A : (b − a) = b; r = tb
B : (c − b) = (−2, 0, −2); r = t(−2, 0, −2) + b
C : (a − c) = −c; r = t(−c) + c = c(1 − t)
Z Z 1 1
1 3
A : (t2 , t2 , t(t + 1)) · (1, 1, 1)dt = t2 + t2 + t2 + t = t3 + t2 =
0 2 0 2
Z Z
(1(−2t + 1), (−2t + 1), (−2t + 1)(2)) · (−2, 0, −2)dt = (−2t + 1)(−2) + (−2)(−2t + 1) =
B: Z
1
= (−2)(−2t + 1)(3) = −6 (−t2 + t) 0 = 0
Z
((1 − t)(−1)(1 − t), −1(1 − t)(−1)(1 − t), (−1)(1 − t)((1 − t) + 1)) · (1, −1, 1) =
Z Z Z
2 2 2 2 2
C : = −(1 − t) + (1 − t) (−1) + (−1)(1 − t) − (1 − t) = −3(1 − 2t + t ) − 1 + t = −3t + 7t − 4 =
1
7t2
−3
= −t3 +
− 4t =
2 0 2
Z
f · ds = 0
Exercise 5. f = (y − z, z − x, x − y)
x = 2 cos φ
x2 + y 2 + z 2 = 4
z = y tan θ y sec θ = 2 sin φ
x2 + y 2 sec2 θ = 4
z = 2 sin φ sin θ
2 sin φ 2 sin φ
f= − 2 sin φ sin θ, 2 sin φ sin θ − 2 cos φ, 2 cos φ −
sec θ sec θ
ds 2 cos φ
= (−2 sin φ, , 2 cos φ sin θ)
dφ sec θ
Z Z
f · ds = 4(sin φ(cos θ − sin θ)(− sin φ) + (sin φ sin θ − cos φ)(cos φ cos θ) + (cos φ − sin φ cos θ) cos φ sin θ)dφ =
Z Z
2 2
= 4 (− sin φ(cos θ − sin θ) + − cos φ(cos θ − sin θ) = 4 (− cos θ + sin θ)dφ = 8π(sin θ − cos θ)
a a
x−
= cos φ
2 2
x2 + y 2 + z 2 = a2 a 2
a 2
a
Exercise 6. f = (y 2 , z 2 , x2 ). 2 2
=⇒ x − 2 + y2 = 2 .
y = sin φ
x + y = ax 2 a
z = a2 − a
2
(cos φ + 1)
2
a2
a 2 a 2
f= sin2 φ,
(1 − cos φ), (cos φ + 1)2
2 2 2
a2
0 −a a
r = sin φ, cos φ, sin φ/z
2 2 4
a 3 a3 a 4 (cos φ + 1)2 sin φ
f · r0 = sin3 φ + (cos φ − cos2 φ) +
2 4 2 z
2π
1
R R
3 2
sin φ = sin φ(1 − cos φ) −→ − cos φ + cos3 φ −−
0
→0
3
Z 2π
cos φ = 0
0
Z 2π
cos2 φ = π
0
ap ap a
z ≥ 0, =⇒ z = 1 + cos φ = 2 cos2 φ/2 = √ | cos φ/2| and so
2 2 2
110
2 2 2
a2 a2 a2 z2 (a2 − z 2 )2
a 4
a
(1 + cos φ)2 =
+ cos φ = + − = so that
2 4 4 4 4 2 22
a (1 + cos φ)2 sin φ (a2 − z 2 )2 sin φ (a4 − 2a2 z 2 + z 4 ) sin φ
Z 4 Z Z
= =
2 z z z
Z Z π Z 2π
π 2π
sin φ/2 cos φ/2/| cos φ/2| = sin φ/2 − sin φ/2 = −2 cos φ/2|0 + 2 cos φ/2|π = (−2)(−1) + 2(−1) = 0
0 π
Z Z 2π 2π
a p a 3/2 2
z sin φ = √ 1 − cos φ sin φ = √ (1 − cos φ) =0
0 2 2 3 0
Z Z 2π 3 3
2π
a p a 2
z 3 sin φ = √ 1 − cos φ sin φ = 3/2 (1 − cos φ)5/2 = 0
0 2 2 5 0
So we finally get
Z
f · r0 dφ = −a3 (π)/4
a3 π
Since we had gone counterclockwise, the exercise asked for the clockwise direction, so reverse the sign to get
R 4
Exercise 7. C (x + y)ds.
a = (0, 0) A: b−a=b r = bt + a = bt |r0 | = |b| = 1 (x + y) = t
√
b = (1, 0) B : c − b = (−1, 1) r = (−1, 1)t + (1, 0) |r0 | = 2 (x + y) = −t + 1 + t = 1
c = (0, 1) C : a − c = −c r = −ct + c = c(1 − t) 0
|r | = 1 (x + y) = 1 − t
Z
1 2 √ 1 √
(x + y)ds = t + t 2 + t − t2 = 2 + 1
C 2 2
y 2 dx
R
Exercise 8. C
α(t) = (a(t − sin t), a(1 − cos t)), 0 ≤ t ≤ 2π
α0 = (a(1 − c(t)), a(s(t)))
α02 = a2 (1 − 2c + c2 + s2 ) = a2 2(1 − c)
Z √ √ √ Z
a2 (1 − 2c + c2 )(a 2 1 − c)dt = a3 2 (1 − c)5/2 dt =
5/2
√ Z 2π Z 2π
t t
= a3 2 2 sin2 dt = a3 21/2 25/2 sin5 dt
0 2 0 2
cos (2t) = cos2 t − sin2 t
since or 2 sin2 t = 1 − cos (2t)
= 1 − 2 sin2 t
Now
5 t 2 t 2 t 2 t 4 t t
sin = (1 − cos ) sin = 1 − 2 cos + cos sin
2 2 2 2 2 2
So
2π 2π
−2 4 −2
Z
t 2π 32
sin5 dt = (−2 cos (t/2))|0 + (4)(1/3) cos3 (t/2) + cos5 (t/2) = (−2) (−2) + +
=
0 2 5 0 3 5 15
Z
32 256 3
y 2 ds = 8a3 = a
C 15 15
Exercise 9. C (x2 + y 2 )ds where C has the vector equation α(t) = (a(cos t + t sin t), a(sin t − t cos t)), 0 ≤ t ≤ 2π
R
α0 (t) = (a((−s + s + tc), c − c + ts)) = a(tc, ts) = at(c, s) ka0 kat
x2 + y 2 = a2 (c2 + 2tcs + t2 s2 + s2 − 2tsc + t2 c2 ) = a2 (1 + t2 )
(2π)2 (2π)3
Z Z
1 1
a2 (1 + t2 )atdt = a3 (t + t3 )dt = a3 (2π)2 + (2π)4 = a3 1+
C 2 4 2 2
R
Exercise 10. C
zdx α(t) = (t cos t, t sin t, t)
111
α0 = (cos t + −t sin t, sin t + t cos t, 1) |α0 |2 = c2 − 2tsc + t2 s2 + s2 + 2sct + t2 c2 + 1 = 2 + t2
t0
(2 + t20 )3/2 − 23/2
Z p
1
(2 + t2 )3/2 =
t 2 + t2 dt =
3 0 3
Exercise 11.
(1)
z=0
r = a(cos t, sin t) 1
Z π
−a
π
2a
y= a sin tadt = cos t =
r0 = a(− sin t, cos t) t ∈ [0, π] πa 0 π 0 π
0 0 Z π
kα (t)k = kr (t)k = a 1
x= a cos tadt = 0
πa 0
(2)
π π
a2 M
Z
1 − cos 2t
Z
M 2 2 1
a sin t(a)dt = dt = M a2
πa 0 π 0 2 2
r = a(cos t, sin t)
Exercise 12. ρ = |x| + |y| r0 = a(−s(t), c(t)) t ∈ [0, 2π]
0 0
kα (t)k = kr (t)k = a
Z π/2 π/2
(a cos t + a sin t)adt = a2 (sin t − cos t) 0 = a2 (1 − (−1)) = 2a2
Z0 π
π
(−a cos t + a sin t)adt = a2 (− sin t − cos t) π/2 = −a2 (−1 + (−1)) = 2a2
π/2
Z 3π/2
=⇒ M = 8a2
3π/2
(−a cos t − a sin t)adt = −a2 (sin t − cos t) π = −a2 (−1 − (−(−1))) = 2a2
π
Z 2π 2π
(a cos t − a sin t)adt = a2 (sin t + cos t) 3π/2 = a2 (1 + 1) = 2a2
3π/2
π/2 π/2 π/2
sin3 t
Z Z
2 2 4 2 2 4 1 3
= a4
(a sin t)(a cos t + a sin t)adt = a sin t cos t + sin t(1 − cos t)dt = a + − cos t + cos t
0 0 3 3 0
Z π Z π
(a2 sin2 t)(−a cos t + a sin t)adt = a4 (− sin2 t cos t + sin t(1 − cos2 t))dt =
π/2 π/2
π
− sin3 t
4 1 3
= a4
=a + − cos t + cos t
3 3 π/2
Z 3π/2 3π/2
− sin3 t
2 2 4 1 3 4 −1 1
= a4
(−a sin t)(a cos t + a sin t)adt = −a + − cos t + cos t = −a + −1 +
π 3 3 π 3 3
Z 2π 3 2π
2 2 4 sin t 1 3
= a4
(a sin t)(a cos t − a sin t)adt = a + cos t − cos t
3π/2 3 3 3π/2
=⇒ I = 4a4
Exercise 13. Notice that we have the plane cut through the center of the sphere: like a conic section, we obtain a circle with
perpendicular √1 (1, 1, 1).
3
2x + 2yyx + y + xyx = 0
x2 + y 2 + z 2 = 1 21 2
=⇒ x + y + xy = =⇒ −y − 2x
x+y+z =0 2 yx =
x + 2y
112
r = (x, y, z)
−y − 2x −y − 2x (y + 2x) x − y
rx = 1, , −1 − = 1, − ,
x + 2y x + 2y x + 2y x + 2y
(y + 2x)2 (x − y)2
3
rx2 = 1 + 2
+ 2
=
(x + 2y) (x + 2y) 2 − 3x2
√
3
|rx | = √
2 − 3x2
q
I guess that x ranges between ± 23 .
p p
We’re given that the mass density is x2 . If we go around the circle on one branch, one p from − 2/3 to 2/3 in
p semicircle,
x, and then around in the same direction on the other branch, other semicircle, from 2/3 to − 2/3 in x, then we calculate
for this branch the same number. So do the calculation for one semicircle.
q q q
Z
x 2 Z 3 2
2x
Z π/2 32
23 sin2 θ 23 cos θdθ r
√ dx = r 2 = p = 3
2 − 3x2 q
3 −π/2 1 − sin2 θ x = sin θ
1− 2
2x where r
π/2 π/2 3
dx = cos θdθ
θ − sin 2θ/2
Z
2 2 2
= sin2 θdθ = = π/3
−π/2 3 3 2
−π/2
2π
=⇒ M =
3
√
Exercise 14. ??? (work on it) Given x2 + y 2 = z 2 , y 2 = x, (0, 0, 0) to (1, 1, 2), curve is parametrized s.t.
p
α = α(y) = (y 2 , y, y 1 + y 2 )
We want z-coordinate of the centroid for a uniform wire.
Consider mass on infinitesimal segment λds.
Weight each λds with corresponding z-coordinate: zλdx
2
Now α0 (y) = (2y, 1, 1 + y 2 + √ y 2 ).
p
1+y
y4
kα0 (y)k2 = 7y 2 + 2 +
1 + y2
ds = kα0 (y)kdy
p
z = y 1 + y2
Then
√
Z Z Z r
p 9 2 17
zds = y 8y 4 + 9y 2 + 2dy = y2 2 (y 2 + ) − 2 dy
16 16
u = y 2 + 9/16 y = 0 =⇒ u = 9/16
du = 2ydy y = 1 =⇒ u = 25/16
Now Z p
1 p 2 p
u2 + β = (u u + β + β ln (u + u2 + β))
2
So then
√ !!
√ Z √ 25 √
25/16
r
−17
Z
17 9 8 25 + 4 38
zds = 2 u2 − 2 du = 2/2 4 38 − + ln
9/16 16 162 16 16 16 17
2
25 − 17 = 608 = 38 ∗ 16
since
81 − 17 = 64
R √8y4 +9y2 +2
√ 2
R
However, M = λ ds = λ dy and we need to divide by M .
y +1
Exercise 15. Given r = (a cos t, a sin t, bt), recall that
Z 2π
p p 8
M= a2 + b2 (a2 + b2 t2 )dt = a2 + b2 (2πa2 + π 3 b2 )
0 3
113
Z p Z 2π p Z 2π
xM = x(x2 + y 2 + z 2 )ds = a2 + b2 a cos t(a2 + b2 t2 )dt = a2 + b2 ab2 t2 cos tdt =
C 0 0
p 2π p p
= a2 + b2 ab2 (t2 sin t + 2t cos t − 2 sin t) 0 = a2 + b2 ab2 (2(2π)) = 4π a2 + b2 ab2
Z p Z 2π
yM = y(x2 + y 2 + z 2 )ds = a2 + b2 a sin t(a2 + b2 t2 )dt =
C 0
p 2π p
= a2 + b2 ab2 2
−t cos t + 2t sin t + 2 cos t 0 = a2 + b2 ab2 (−(2π)2 )
Exercise 16.
Z p Z
Ix = 2 2 2
(y + z )(x + y + z )ds = 2 2
(a2 sin2 t + b2 t2 )(a2 + b2 t2 )dt =
a2 +b2
C C
1 − cos 2t 1 − cos (2t)
Z Z
2 4 2 2 2 2 2 2 2 2 4 4 2 4
= ρ0 (a sin t + a b t + a b t sin t + b t )dt = ρ0 a + a2 b2 t2 + a2 b2 t2 + b4 t4 dt =
C C 2 2
4 2 2 3 2 2 3 4 5 2 2
(2π)3 a2 b2 (2π)5 b4 π(a2 b2 )
a (2π) a b (2π) a b (2π) b (2π) πa b p
= ρ20 + + + − = a2 + b2 πa4 + + −
2 3 2(3) 5 2 2 5 2
Z Z
Iy = (x2 + z 2 )(x2 + y 2 + z 2 )ds = ρ20 (a2 cos2 t + b2 t2 )(a2 + b2 t2 )dt =
C
Z Z C
2 4 2 2 2 2 2 2 2 2 4 4 2 4 1 + cos 2t 2 2 2 2 2 2 1 + cos (2t)
= ρ0 (a cos t + a b t + a b t cos t + b t )dt = ρ0 a +a b t +a b t + b4 t4 dt =
C C 2 2
4
a (2π) a2 b2 (2π)3 a2 b2 (2π)3 a2 b2 π b4 (2π)5 (2π)3 a2 b2 (2π)5 b4 πa2 b2
p
2 2 2 4
= ρ0 + + + + = a + b πa + + +
2 3 2(3) 2 5 2 5 2
10.13 E XERCISES - O PEN CONNECTED SETS . I NDEPENDENCE OF THE PATH . T HE SECOND FUNDAMENTAL THEOREM
OF CALCULUS FOR LINE INTEGRALS . A PPLICATIONS TO MECHANICS .
Exercise 1. Recall the lesson of the preceding sections.
Let S be an open set in Rn . The set S is called connected if every pair of points in S can be joined by a piecewise smooth
path whose graph lies in S. That is, for every pair of points a and b in S there is a piecewise smooth path α defined on an
interval [a, b] such that α(t) ∈ S for each t ∈ [a, b] with α(a) = a, α(b) = b.
(1) S = {(x, y)|x2 + y 2 ≥ 0} connected.
Consider (x1 , y1 ) = (r1 , θ1 ) (x2 , y2 ) = (r2 , θ2 )
Suppose r2 > r1 . Let α1 = (r cos θ2 , r sin θ2 ) s.t. r1 ≤ r ≤ r2
Then consider α2 = (r1 cos θ, r1 sin θ) s.t. θ : θ2 → θ1
(2) S = {(x, y)|x2 + y 2 > 0} connected.
See part(a), with α1 , α2
(3) S = {(x, y)|x2 + y 2 < 1} connected.
See part(a), with α1 , α2 but with r1 , r2 < 1
(4) S = {(x, y)|1 < x2 + y 2 < 2} connected.
See part (a), with α1 , α2 , but with 1 < r1 , r2 < 2
Exercise 2. f = (P, Q) = ∂ϕ ,
∂x ∂y
∂ϕ
∂P ∂Q ∂ϕ ∂2ϕ ∂Q
Since ∂y , ∂x continuous, then by Apostol Vol. 2, Thm. 8.12, ∂P ∂
∂y = ∂y ∂x = ∂x∂y = ∂x
Exercise 3.
∂P
(1) ∂y =1
∂Q
∂x = −1 f (x, y) = yi − xj
Consider a square contour lying on x and y axes.
Z (1,0) Z (1,1)
f dx = 0 (−1)dy = −1
(0,0) (1,0)
Z (0,1) Z (0,0)
1dy = −1 0=0
(1,1) (0,1)
R
So then C
f · ds = −2
114
∂Q
(2) f (x, y) = yi + (xy − x)j ∂P
∂y = 1 ∂x = y − 1
Z (1,0) Z (1,1)
f dx = 0 (y − 1)dy = 1/2 − 1 = −1/2
(0,0) (1,0)
Z (0,1) Z (0,0)
1dx = −1 = −1 0=0
(1,1) (0,1)
R
C
f · ds = −3/2
Exercise 4. f (x, y, z) = P (x, y, z)i + Q(x, y, z)j + R(x, y, z)k
We’re given that
∂y P, ∂z P, ∂x Q, ∂z Q, ∂x P, ∂y R
are cont.
Now
f = ∇ϕ = (P, Q, R) = (∂x ϕ, ∂y ϕ, ∂z ϕ)
∂y P = ∂y ∂x ϕ = ∂x ∂y ϕ = ∂x Q
∂z P = ∂z ∂x ϕ = ∂x ∂z ϕ = ∂x R
∂z Q = ∂z ∂y ϕ = ∂y ∂z ϕ = ∂x R
Exercise 6.
(1) P = y, Q = z.
∂y P = 1 ∂x Q = 0 Not conservative.
(2) Since f · dα = f · α0 (t)dt
α(t) = (cos t, sin t, et ) =⇒ α0 (t) = (− sin t, cos t, et )
f = (y, z, yz).
f · α0 (t) = (− sin2 t + et cos t + sin te2t )
Z π t π
e sin t + et cos t −e2t cos t
2 t 2t 2t
(− sin t + e cos t + sin te )dt = −π + + e sin t + /(5/2) =
0 2 2 0
π 2π
e (−1) − 1 2 −e + 1
= −π + +
2 5 2
11.9 E XERCISES - I NTRODUCTION . PARTITIONS OF RECTANGLES . S TEP FUNCTIONS . T HE DOUBLE INTEGRAL OF A
STEP FUNCTION . T HE DEFINITION OF THE DOUBLE INTEGRAL OF A FUNCTION DEFINED AND BOUNDED ON A
RECTANGLE . U PPER AND LOWER DOUBLE INTEGRALS . E VALUATION OF A DOUBLE INTEGRAL BY REPEATED
ONE - DIMENSIONAL INTEGRATION . G EOMETRIC INTERPRETATION OF THE DOUBLE INTEGRAL AS A VOLUME .
W ORKED EXAMPLES .
Exercise 1.
1
y y2
Z
A(y) = yx(x + y)dx = +
0 3 2
Z 1
1 1 1
A(y)dy = + =
0 6 6 3
Exercise 2.
Z 1
1
A(y) = dx(x3 + 3x2 y + y 3 ) = + y + y3
0 4
Z 1
1 1 1
A(y) = + + = 1
0 4 2 4
Exercise 3.
1
√ √
Z
1 3
( y + x − 3xy 2 )dx = y + − y 2
A(y) =
0 2 2
3
Z 3
2 3/2 y y 3 √ √
√ 1 3 −2 3 1 27 1 38
y + − y2 = y + − =2 3+ + − − + = 2 3−
1 2 2 3 2 2 1 3 2 2 2 2 3
Exercise 4.
115
Z π
π
A(y) = sin2 x sin2 ydx = sin2 y
0 2
π π π2
Z
A(y)dy = =
2 2 4
Exercise 5.
Z π/2 π
π/2
A(y) = sin (x + y)dx = − cos (x + y)|0 = cos y − cos + y = cos y + sin y
0 2
Z π/2
π/2 π/2
A(y)dy = + sin y|0 − cos y|0 = 2
0
RR
Exercise 6. Split the integral up into 4 parts. Given Q | cos (x + y)|dxdy where Q = [0, π] × [0, π]
Z −x+π/2
A(x) = cos (x + y)dy = 1 − sin x
0
Z π =⇒ 2
π/2−x
A(x) = − cos (x + y)dy = sin (x + y)|π = 1 − sin (x + π) = 1 + sin x
−x+π/2
Z −x+3π/2
0
A(x) = − cos (x + y)dy = sin (x + y)|3π/2−x = sin x − (−1) = 1 + sin x
0
Z π =⇒ 2
π
A(x) = cos (x + y)dy = sin (x + y)|−x+3π/2 = sin (x + π) − (−1) = 1 − sin x
−x+3π/2
Z π
A(x) = 2π
0
RR
Exercise 7. Q
f (x + y)dxdy and Q = [0, 2] × [0, 2], f (t) greatest integer ≤ t
y<1
Z 2−y Z 2
A(y) = 1dx + 2dx = 2 − y − (1 − y) + 2(2 − (2 − y)) = 2y + 1
1−y 2−y
y>1
Z 2−y Z 3−y Z 2
A(y) = dx + 2dx = 3dy = 2 − y + 2(3 − y − (2 − y)) + 3(2 − (3 − y)) = 2y + 1
0 2−y 3−y
Z 2
(2y + 1) = 6
0
y −3 etx/y dxdy, and Q = [0, t] × [1, t], t > 0
RR
Exercise 8. Q
t t 2
y −3 etx/y y −3 et /y y −2
Z
y −3 etx/y dx = = −
0 t/y
0 t/y t
Z t 2
t 2
2 −et /y −et et
y −2 et /y dy = = +
t2 t2 t2
1
1
Z t t
−1 −1 1
y −2 /tdy = = 2 +
1 yt 1 t t
2
et et −1 1
=⇒ − 3
+ 3 + + 2
t t t t
Exercise 9. Q rectangle, Q = [a, b] × [c, d].
! !Z
ZZ Z Z b Z b d
f (x)g(y)dxdy = f (x)dx g(y)dy = f (x)dx g(y)dy
Q a a c
Rb
Assume a
f (x)dx = A
( exists.
1−x−y if x + y ≤ 1
Exercise 10. f (x, y) =
0 otherwise
116
Z 1−y
1 1 1
A(y) = (1 − x − y)dx = 1 − y − (1 − 2y + y 2 ) − y(1 − y) = − y + y 2
0 2 2 2
Z 1
1 1 1
A(y)dy = − + = 1/6
0 2 2 6
Indeed, vol. of tetrahedron = 13 Bh = 13 12 (1)(1) = 1/6
Exercise 11. If x < √1
2
Z 2x2
1 4 4 3 2 3
A(x) = (x + y)dy = xx + (4x − x ) = x 1 + x
x2 2 2
√
Z 1/ 2 1/√2 √ ! √
1 4 3 5 1 1 3 1 2 1 3 2
=⇒ A(x)dx x + x = + = +
0 4 10 0 4 4 10 4 2 16 80
√
When x > 1/ 2,
1
1 x4
Z
1
A(x) = (x + y)dy = (x)(1 − x2 ) + (1 − x4 ) = x − x3 + −
x2 2 2 2
1
√
1 x4 −16 2
Z
3 1 1 1 1 1 1 1 1 84
√ (x − x + − )dx = (1 − ) − (1 − ) + (1 − √ ) − (1 − √ ) = +
1/ 2 2 2 2 2 4 4 2 2 10 4 2 160 80
√
21 2
So we get −
40 5 (
x2 + y 2 if x2 + y 2 ≤ 1
Exercise 12. f (x, y) =
0 otherwise
√
Z 1−x2
2
p 1 p p 4 p 2p
A(x) = √ (x + y 2 )dy = 2x2
1 − x2 + ( 1 − x2 )3 − (− 1 − x2 )3 = x2 1 − x2 + 1 − x2
− 1=x2 3 3 3
Now
Z 1 p Z π/2 Z π/2 2
1 π/2 1 − cos 4θ
Z
2 2 2 2 sin 2θ
x 1−x = sin θ cos θdθ = dθ = = π/8
−1 −π/2 −π/2 2 4 −π/2 2
Z p Z π/2
2
1 − x dx = cos2 θdθ = π/2
−π/2
4 π 2 π
=⇒ + = π/2
3 8 3 2
Exercise 13. Split the integral up into 2 parts.
Z y y
(x + y)−2 dx = − (x + y)−1 1 = −(2y)−1 + (y + 1)−1
1
Z 2Z y
−1 3
(x + y)−2 dxdy = ln 2 + ln
1 1 2 2
Z 2 −1
−2 −1 2
−1 3y
(x + y) dx = −(x + y) y/2 = −(2 + y) +
y/2 2
Z 4Z 2
4 2 2
(x + y)−2 dxdy = − ln (2 + y)|2 + ln 2 = − ln 6 + 2 ln 2 + ln 2
2 y/2 3 3
ln 2
Add the two up to get (it may help to remember that ln 3 + − ln 6 = ln 3 + ln 1/6 = ln 1/2 = − ln 2)
6 (
1 if x = y
Exercise 14. Q = [0, 1] × [0, 1] and f (x, y) =
0 if x 6= y
Let D = {(x, y)|x = y}, D = “diagonal” of Q.
D has content zero as it’s the graph of a continuous function y = x, 0 ≤ x ≤ 1.
117
f discontinuous only on D.
so f continuous on D, with D of content zero, so f integrable on R.
ZZ
f = 0 since f = 0 ∀ (x, y) ∈ Q − D
Q−D
11.15 E XERCISES - I NTEGRABILITY OF CONTINUOUS FUNCTIONS . I NTEGRABILITY OF BOUNDED FUNCTIONS WITH
DISCONTINUITIES . D OUBLE INTEGRALS EXTENDED OVER MORE GENERAL REGIONS . A PPLICATIONS TO AREA AND
VOLUME . W ORKED EXAMPLES .
Exercise 1. Z x
x
x cos (x + y)dy = x sin (x + y)|0 = x(sin 2x − sin x)
0
π π
−3π
Z
x cos 2x sin 2x π
dxx(sin 2x − sin x) = + + x cos x − sin x = −π =
0 −2 4 0 −2 2
Exercise 2.
Z x+1
(1 + x) sin ydy = (1 + x)(1 − cos (x + 1))
0
Z 1
1 1 3
(1 + x) − (1 + x) cos (x + 1) = 1 + − (1 + x) sin (x + 1) − cos (x + 1)|0 = − 2 sin 2 − cos 2 + sin 1 + cos 1
0 2 2
RR x+y
Exercise 3. S e dxdy where S = {(x, y)||x| + |y| ≤ 1}
Z 1−x Z 1+x
e x+y x 1−x
dy = e (e −e −1+x 1
)=e −e −1+2x
ex+y dy = ex (e1+x − e−x−1 ) = e1+2x − e−1
−1+x −x−1
0
1 −1 −1 e1 −3e−1 e1
Z Z
e e e
(e+1
− e−1+2x )dx = e − (e2 − 1) = + (e1+2x − e−1 )dx = (1 − e−2 ) − e−1 (1) = +
0 2 2 2 −1 2 2 2
=⇒ e − e−1
xy = 1
xy = 2 1 2
x2 y 2 dxdy and
RR
Exercise 4. S y= y =
y=x x x
y = 4x
√ √
I : (1/2, 2) → (1/ 2, 2 2)
Z 4x
1 1
64x3 − 3 x2 y 2 dy = x2
1/x 3 x
√
Z 1/√2 1/ 2
1 64x6
1 1 1 4 1 1 1 1
64x5 −
dx = − ln x = − − ln √ − ln =
3 1/2 x 3 6 1/2 3 3 6 3 2 2
7 1 1
= − ln 2
18 3 2
Z 2/x
2 2 21 8 1 71
x y dy = x − 3 =
√ √ 1/x 3 x 3 x 3x
II : (1/ 2, 2 2) → (1, 1) Z 1
71 7 1 7
√ = − ln √ = ln 2
1/ 2 3 x 3 2 6
Z 2/x
1 8 81 1 5
x2 y 2 dy = x2 − x3 = − x
√ √ x 3 x3 3x 3
III : (1, 1) → ( 2, 2) √ √2
2
8 √
Z
81 1 5 1 6 4 7
− x = ln 2 − x = ln 2 −
1 3x 3 3 18 1 3 18
4 7 7 7 1 7
=⇒ ln 2 − + ln 2 + − ln 2 = ln 2
3 18 6 18 6 3
(x2 − y 2 )dxdy.
RR
Exercise 5. S
118
Z sin x
1 1
(x2 − y 2 )dy = x2 sin x − sin3 x = x2 sin x − (1 − cos2 x) sin x
0 3 3
Z
x2 sin x = −x2 cos x + 2x sin x + 2 cos x
π
cos3 x
−40
ZZ
1 2 2
(x2 − y 2 )dxdy = −π 2 (−1) + 2((−1) − 1) + (−1 − 1) + = π 2 + −4 − + = + π2
S 3 −9
0 3 9 9
6−x−2y
Exercise 6. x + 2y + 3z = 6 =⇒ z = 3
3− x
x2 x2
6 − x − 2y
Z 2 1 x x x 2 1
dy = 6 3− −x 3− − 3− = 18 − 3x − 3x + − 9 − 3x + =
0 3 3 2 2 2 3 2 4
x2
1
= 9 − 3x +
3 4
Z 6 2
2
1 x 1 3(6) 1 3 1
9 − 3x + = 9(6) − + 6 = (6)(9 − 9 + 3) = 6
0 3 4 3 2 12 3
Indeed, 31 BH = 13 (9)(2) = 6
Exercise 7.
Z x
1 1 4
(x2 − y 2 )dy = x2 (x − (−x)) − (x3 − (−x)3 ) = 2x3 − (x3 + x3 ) = x3
−x 3 3 3
Z 3 3
4 3 1 1 80
x dx = x4 = (81 − 1) =
1 3 3 1 3 3
Exercise 8.
(1)
Z 1
1 2
(x2 + y 2 )dy = x2 (2) + (13 − (−1)3 ) = 2x2 +
−1 3 3
Z 1
2 2 2
2x2 + = (2) + (2) = 8/3
−1 3 3 3
(2) f (x, y) = 3x + y. S = {(x, y)|4x2 + 9y 2 ≤ 36, x > 0, y > 0}. Thus, y 2 ≤ 4 − 94 x2
Z 2√1−(x/3)2 r x 2 1 x 2 r x 2 x 2
(3x + y)dy = 3x2 1 − + 4(1 − ) = 6x 1 − + 2(1 − )
0 3 2 3 3 3
Z 3 r 3
x 2 x2 x2 3/2 2
6x 1 − + 2(1 − ) = −18(1 − ) + 2(3) − (27) = 22
0 3 9 9 0 27
x2 3/2 0 3 x2 −2x −x x2
since ((1 − ) ) = (1 − )1/2 ( )= (1 − )1/2
9 2 9 9 3 9
(3)
√
Z 16−x2 p p p
√ (y + 2x + 20)dy = (2x + 20)(2 16 − x2 ) = 4x 16 − x2 + 40 16 − x2
− 16−x2
Z 4 Z π
p p 640
4x 16 − x2 + 40 16 − x2 = (16 cos θ4 sin θ + 40(4) sin θ)4 sin θdθ = π = 320π
−4 0 2
x = 4 cos θ
since
dx = −4 sin θdθ
R1 Ry R 1 R 1 R 2 R 2y R 4 R √x
Exercise 9. 0 0 f (x, y)dx dy = 0 x
f (x, y)dy dx Exercise 10. 0 y 2 f (x, y)dx dy = 0 x/2
f (x, y)dy dx
R4 R2 R 2 R y2
Exercise 11. 1 √ f (x, y)dy dx = f (x, y)dx dy
x 0 0
√ R 1 R √1−y2 +1
R 2 R 2x−x2 R 2 R 2−x
Exercise 12. 1 2−x
f (x, y)dy dx = 0 2−y
f (x, y)dx dy Exercise 13. −6 2
(x −4)/4
f (x, y)dy dx =
R 0 R √4y+4 R 8 R 2−y R e R log x R1 Re
−1
√ f (x, y)dx dy+ 0 −√4y+4 f (x, y)dx dy Exercise 14. 1 0
− 4y+4
f (x, y)dy dx = 0 ey f (x, y)dx dy
R 1 R √1−y R 0 R √1−y2
R 1 R 1−x2
Exercise 15. −1 −√1−x2 f (x, y)dy dx = 0 √
− 1−y
f (x, y)dx dy + −1
√ 2
f (x, y)dx dy Exercise 16.
− 1−y
119
R 1 R x2 R 1 R y1/3
0 x3
f (x, y)dy dx = 0 √y f (x, y)dx dy
Exercise 17. Consider that sin (π − x) = −1 sin (−x) = sin x = y. This way, we get the “branch” of values for π/2 < x < π
and 0 < y < 1.
Z π Z ! !
sin x Z 1 Z π−arcsin y Z 0 Z π
f (x, y)dy dx = f (x, y)dx dy + f (x, y)dxdy
0 − sin (x/2) 0 arcsin y −1 2 arcsin (−y)
R 4 R (y−4)/2 R 0 R 4−x2
Exercise 18. 0
√
− 4−y
f (x, y)dx dy = −2 2x+4 f (x, y)dy dx
Exercise 19.
Z 1 Z y Z 2 Z 2−y Z 1 Z −2x
2 2 2 2 2 2
V = (x + y )dx dy + (x + y )dx dy = (x + y )dy dx
0 0 1 0 0 x
Z 2−x
1 1
(x2 + y 2 )dy = x2 (2 − x − x) + ((2 − x)3 − x3 ) = x2 (2 − 2x) + (8 − 4x(3) + 3(2)x2 − x3 − x3 ) =
x 3 3
8 2x3 −8 3 8
= 2x2 − 2x3 + − 4x + 2x2 − = x + 4x2 − 4x +
3 3 3 3
−2 4
R1
0 8 4
−−→ + −2+ =
3 3 3 3
Exercise 21.
R 2 R x3 R 8 R 8 R 8 R y
(1) V = 1 x
f (x, y)dy dx + 2 x f (x, y)dy dx = 1 y1/3 f (x, y)dx dy
(2)
R1 R
x −y 2
Exercise 22. I = −1/2 0
e dy dx
Z 1 Z x Z 1 Z x Z 0 Z x
2 2 2
e−y dy dx = e−y dy dx + e−y dy dx
−1/2 0 0 0 −1/2 0
Z 1 Z x Z 1 Z 1 Z 1 Z 1
2 2 2 2
e−y dy dx = e−y dx dy = e−y (1 − y)dy = A + −ye−y dy =
0 0 0 y 0 0
! 1 2
e−y e−1 1
=A+ =A+ −
2 2 2
0
z = −x
since
dz = −dx
Z 0 Z x Z 0 Z −x Z 1/2 Z −x Z 1/2 Z 0
2 2 2 2
e−y dy dx = − e−y dy dx = e−y dy dx = − e−y dy dx =
−1/2 0 1/2 0 0 0 0 −x
!
Z 0 Z 1/2 Z 0
2 2 1
=− e−y dx dy = − e−y − (−y) dy =
−1/2 −y −1/2 2
2
! 0
0 0 0
−1 −e−y
Z Z Z
−y 2 −y 2 1 −y 2
= e dy + − ye dy = e (+dy) + − =
2 2 2
−1/2 −1/2 1/2
−1/2
1/2
−1 1 −e−1/4
Z
2
= e−y dy + +
2 0 2 2
=⇒ I = 2A + e−1 − 1 + −B + 1 − e−1/4 = 2A − B + e−1 − e−1/4
Exercise 23.
(1) Suppose S is a type I region, without loss of generality.
Use the geometry of similar triangles. Thus, observe that the cross-sectional area is a projection of plane region S
(from the geometry of similar triangles). Now express this mathematically.
Z b Z φ2 (x) Z b Z φ2 x
A= dx dy = dydx
a φ1 (x) a φ1 (x)
120
Z bt/h Z φ2 (x)( ht ) Z bt/h Z φ2 2 Z b Z φ2 2
t t t
dx dy = dx dY = dX dY = A
at/h φ1 (x)( ht ) at/h φ1 h h a φ1 h
Rh t2 1
(2) 0 h2
A = hA
3
Exercise 24.
Z 1 Z a Z 1
m(a−x)
e f (x)dy dx = em(a−x) f (x)(a − x)dx
0 x 0
11.18 E XERCISES - F URTHER APPLICATIONS OF DOUBLE INTEGRALS , T WO THEOREMS OF PAPPUS
Exercise 1.
Z 1 Z 2−x Z 1 1
1 3 1 4 1 12 2
xdydx = x((2 − x) − x ) = x − x − x = 1 − 4 − (1 + 8) − (1 − 16) = −9/4
−2 x2 −2 3 4 −2 3 4
Z 1 Z 2−x Z 1 Z 1 1 1
1 2 4 1 2 4 2 1
1 3 1 5
ydydx = ((2 − x) − x )dx = (4 − 4x + x − x )dx = 2(1 + 2) − x −2 + x − x =
−2 x2 −2 2 −2 2 8 −2 10 −2
1 1 72
= 6 + (1 + 8) − (1 + 32) =
6 10 10
Z 1 Z 2−x Z 1
1 1
dydx = (2 − x − x2 ) = 2(1 + 2) − (1 − 4) − (1 − (−8)) = 9/2
2 x 2 −2 2 3
x = −1/2, y = 8/5
2
y =x+3 (−3, 0), (1, ±2)
Exercise 2. 2
y =5−x (1, ±2), (5, 0)
2
! 2
Z 2 Z 5−y Z 2 Z 2
2 3 2
dx dy = 5 − y 2 − (y 2 − 3)dy = 8 − 2y 2 dy = 8(4) − y = 32 − (8 − (−8)) =
−2 y 2 −3 −2 −2 3 −2 3
96 − 32 64
= =
3 3
5−y 2
! 2 !
2 2
1 2
Z Z Z Z
1 1 4 3
xdx dy = (25 − 10y 2 + y 4 − (y 4 − 6y 2 + 9))dy = (16 − 4y 2 )dy = 16(4) − y =
−2 y 2 −3 −2 2 2 −2 2 3 −2
1 1 192 − 64 1 64
= (64 − 4/3(16)) = = (128) =
2 2 3 6 3
Z 2 Z 5−y2 Z 2
ydxdy = 8y − 2y 3 = 0
−2 y 2 −3 −2
x = 1, y=0
x − 2y + 8 = 0 x = −2
Exercise 3.
x + 3y + 5 = 0 x=4
x+8 4
4 Z 4 Z 4
5x2
Z Z
5+x 2 3x + 24 + 10 + 2x 5x + 34 17 5
dydx = dx = dx = + (6) = (12)+34 = −39
−2 −3 −2 6 −2 6 12 −2 3
12
Z 4 2 4
5x + 34x 1 5 3 1 5 1
x + 17x2 =
xA = dx = (64 + 8) + 17(16 − 4) = (120 + 17(12)) = 54
−2 6 6 3 −2 6 3 6
Z 4 2 2
Z 4
dx 9x + 9(16x) + 9(64) − 100 − 40x − 4x2
2
1 (x + 16x + 64) 25 + 10x + x
yA = dx − = =
−2 2 4 9 −2 2 36
Z 4 4
1 1 5 3
dx(5x2 + 18(8x) − 8(5x) + 9(4)(16) − 4(25)) = x + 52x2 + 119(4x) =
=
2(36) −2 36 3 −2
1 5
= (64 + 8) + 52(16 − 4) + 119(4)(6) = 50
2(36) 3
50 18
y= x=
39 13
121
Exercise 4. y = sin2 x; y = 0; 0 ≤ x ≤ π
π sin2 x π π
1 − cos 2x
Z Z Z Z
π
dydx = sin2 xdx =
dx =
0 0 0 0 2 2
Z π Z sin2 x Z π Z π π
π2 π2
2 1 − cos 2x 1 x sin 2x cos 2x
xdydx = x sin xdx = x dx = − + =
0 0 0 0 2 4 2 2 4
0 4
Z π Z sin2 x Z π Z π 2 Z π
1 1 1 − cos 2x 1 1 π 3π
ydydx = sin4 xdx = dx = 1 − 2 cos 2x + cos2 2x = (π + ) =
0 0 0 2 2 0 2 8 0 8 2 16
3π
3
x = π 2 /4/π/2 = π/2 y= 16
π =
2 8
Exercise 5.
Z π/4 Z cos x Z π/4
π/4
√
dydx = cos x − sin x = (sin x + cos x)|0 = 2−1
0 sin x 0
π/4 cos x π/4
√ √
π 2
−1
Z Z Z
π/4 π 2 4
xdydx = x(cos x − sin x)dx = (x sin x + cos x + x cos x − sin x)|0 = −1 x= √
0 sin x 0 2 2 2−1
π/4 π/4
1 π/4
Z Z
1 1 1 1/4
(cos2 x − sin2 x)dx =
cos 2xdx = sin 2x = y=√
0 2 2 0 4 0 4 2−1
y = log x
Exercise 6. , 1 ≤ x ≤ a.
y=0
Z a Z log x Z a
a
dydx = log xdx = (x log x − x)|1 = a log a − a + 1
1 0 1
2
a 2a log a−(a2 −1)
a log x a 2
2 2 2
a log a (a − 1)
Z Z Z
x log x x 4
xdydx = xdx log x = − = − =⇒ x =
1 0 1 2 4 1 2 4 a log a − a + 1
Z a Z log x Z a a
1 1 1
(log x)2 dx = (log x)2 x − 2x log x + 2x = (a(log a)2 − 2a log a + 2a − 2) =
ydydx =
1 0 1 2 2 1 2
1 2
= a(log a) − a log a + a − 1
2
((log x)2 x)0 = (log x)2 + 2(log x) 1
a(log a)62 − a log a + a − 1
0
=⇒ y = 2
(x log x − x) = log x a log a − a + 1
√ √ √ √ √
Exercise 7. x + y = 1 or y = 1 − x. x = 0, y = 0. So y = 1 − 2 x + x.
√ 1
Z 1 Z 1−2 x+x
√
Z 1
4 1 4 1 1
(1 − 2 x + x) = (x − x3/2 + x2 ) = 1 − + =
dydx =
0 0 0 3 2 0 3 2 6
Z 1 Z 1−2√x+x Z 1 Z 1
1 √ 1
ydydx = dx (1 − 4 x + 2x + 4x − 4x3/2 + x2 ) = dx (1 − 4x1/2 + 6x − 4x3/2 + x2 ) =
0 0 0 2 0 2
5/2
1
1 8 8x 1 1
= (x − x3/2 + 3x2 − + x3 ) =
2 3 5 3 0 30
=⇒ y = 1/5
√ 1
Z 1Z 1−2 x+x Z 1 √
1 2 2 5/2 1 3 1
xdydx = x(1 − 2 x + x)dx = x −2 x + x =
0 0 0 2 5 3 0 3
=⇒ x = 1/5
2/3 2/3
y 2/3 = 1 − x2/3
Exercise 8. x +y = 1 or and x = 0, y = 0.
y = (1 − x2/3 )3/2
122
u = 1 − x2/3 or x2/3 = 1 − u
−2 −1/3
Since du = 3 x dx
−3 √
du( 1 − u) = dx
2
Z 1 Z (1−x2/3 )3/2 Z 1 Z 0
−3
dydx = (1 − x2/3 )3/2 dx = u3/2 du(1 − u)1/2
0 0 0 1 2
0
−2 −2 3 1/2 −2
Z Z Z
u3/2 (1 − u)1/2 = u3/2 (1 − u)3/2 = u3/2 (1 − u)3/2 − u (1 − u)(1 − u)1/2 =
3 3 2 3
−2 3/2
Z
= u (1 − u)3/2 + (u1/2 (1 − u)1/2 − u3/2 (1 − u)1/2 )
3
−2 3/2
Z Z
=⇒ 2 u3/2 (1 − u)1/2 = u (1 − u)3/2 + u1/2 (1 − u)1/2
3
Z Z Z 2 !1/2 Z
1/2 1/2 2 1/2 1 1 1 p
u (1 − u) = (u − u ) = − −u = 1 − (1 − 2u)2
4 2 2
x = (1 − 2u) x = sin θ
Since and then
dx = −2du dx = cos θdθ
−1
Z p Z Z Z
2
dx p 2
cos θdθ 1 + cos 2θ sin 2θ
1 − (1 − 2u) = 1−x = cos θ = = (θ + )
−2 −2 −4 4 2
−1 −1 p
Z 1
1 1 1/2 1 1p
Z Z Z
u3/2 (1 − u)1/2 du = u (1 − u)1/2 = 1 − (1 − 2u)2 = dx 1 − x2 =
0 2 0 4 0 8 1
−1 −π/2 1 + cos 2θ
Z
= dθ =
8 π/2 2
−π/2
−1 sin 2θ −1 −π π
= θ+ = − = π/16
16 2
π/2 16 2 2
3 π 3π
=⇒ A = =
2 16 32
1 (1−x2/3 )3/2 1 0 1
−3 √
Z Z Z Z Z
2/3 3/2 3
xdydx = x(1 − x ) dx = du 1 − uu3/2 (1 − u)3/2 = du(1 − 2u + u2 )u3/2 =
0 0 0 1 2 2 0
3 1
Z
= du(u3/2 − 2u5/2 + u7/2 ) =
2 0
3 2 2 2 3 2 4 2 8
= −2 + = − + =
2 5 7 9 2 5 7 9 105
=⇒ x = 256π/315
1 (1−x2/3 )3/2 1
1 1
Z Z Z Z
1
ydydx = (1 − x2/3 )3 dx = (1 + 3(−x2/3 ) + 3x4/3 − x2 )dx =
0 0 0 2 2 0
3(1)5/3
1 3 1 1 −9 9 −1 1 −4 9 1 8
= 1 + −3 +3 − = 1+ + + = + − =
2 5 7 3 2 5 7 3 2 5 7 3 105
=⇒ y = 256π/315
R 2 R x(2−x) R2 R x(2−x)
1−y 1
Exercise 9. 0 0 1+x dydx = 0 1+x 0
(1 − y)dydx
1 1 1
(x(2 − x))2 = x2 (4 − 4x + x2 ) = 2x2 − 2x3 + x4
2 2 2
Z x(2−x)
1 1
(1 − y)dy = 2x − x2 − 2x2 + 2x3 − x4 = 2x − 3x2 + 2x3 − x4 =
0 2 2
−1 4 1 3 5 3 5 2 11 2 11 15 −1 3 5 2 11 15
= x − x + x + x − x − x+ x= x + x − x (x + 1) + x
2 2 2 2 2 2 2 2 2 2 2
123
2 x(2−x) 2
−1 3 5 2 11x
Z Z Z
1 15x 1
(1 − y)dydx =
x + x − (x + 1) + =
0 0 1+x 0 2 2 2 2 1+x
Z 2
−1 3 5 2 11x 15x
= dx x + x − + =
0 2 2 2 12(1 + x)
−1 5 11 15 2
20 15 26 −15
= (16) + (8) − (4) + 2 + ln (1 + x)|0 = −2 + − 11 + (2 − ln (3)) = + ln 3
8 6 4 2 3 2 3 2
Exercise 10.
Z a Z b Z a
1 1
(xy)dydx = x b2 dx = a2 b2
0 0 0 2 4
Z a Z b Z a
1 1
(xy 2 )dydx = x b3 dx = a2 b3
0 0 0 3 6
Z a Z b Z a
1 1
x2 ydydx = x2 b2 = b2 a3
0 0 0 2 6
1 2 3 1 2 3
6b a 2 6a b 2
x= 1 2 2 = a y= 1 2 2 = b
4a b
3 4a b
3
Exercise 11.
y = sin2 x
−π ≤ x ≤ π; f (x, y) = 1
y = − sin2 x
sin2 x
! 3
π π π
1 − cos 2x
Z Z Z Z
1 2
Ix = 2
y dy dx = 2 sin6 xdx = =
−π − sin2 x −π 3 3 −π 2
Z π
1 5π
= 1 + −3 cos 2x + 3 cos2 2x + cos3 2x =
12 −π 12
π sin2 x π π
1 − cos 2x
Z Z Z Z
2 2 2 2
Iy = x dydx = x 2 sin xdx = 2 2x dx =
−π − sin2 x −π −π 2
π
2π 3
Z
= x2 − x2 cos 2x = −π
−π 3
x y x y
Exercise 12. a + b = 1, c + b = 1, y = 0, 0 < c < a, b > 0; f (x, y) = 1
! b
b a(1−y/b) b
a3 (1 − y/b)3 − c3 (1 − y/b)3 a3 − c3
Z Z Z
2
y 4 b
Iy = x dx dy = dy = (−b) 1 − /b = (a3 − c3 )
0 c(1−y/b) 0 3 3 b a 12
b
1 3 1 b4
Z
Ix = y 2 (a − c)(1 − y/b)dy = (a − c) b − = (a − c)(b3 )/12
0 3 4 b
Exercise 13. (x − r) + (y − r)2 = r 2 , y = 0, x = 0. 0 ≤ x ≤ r, 0 ≤ y ≤ r. So we want the piece of the graph that is the
2
“lower left-hand corner” of the “complement” of the circle centered at (r, r), with radius r. Be careful about this point.
x − r = r sin θ
since
dx = r cos θdθ
Z r Z r−√r2 −(x−r)2
s 3
r 2 4 0
x−r
Z Z
1 3 dx = r
Ix = y 2 dydx = r 1− 1− (1 − cos θ)3 cos θdθ =
0 0 0 3 r 3 −π/2
0
r4
Z
5π
= c − 3c2 + 3c3 − c4 = r4 1 −
3 −π/2 16
124
Since Z 0
0
c = s|−π/2 = −(−1) = 1
−π/2
Z 0 Z 0
2 1 + cos 2θ π
c = dθ =
−π/2 −π/2 2 4
Z 0 Z 0
1
c(1 − s2 ) = 1 − (0 − (−1)) = 2/3
c3 =
−π/2 −π/2 3
Z 0 Z 0 2 Z 0
1 + cos 2θ 1 1 π π 3π
c4 = dθ = (1 + 2 cos 2θ + cos2 2θ)dθ = + =
−π/2 −π/2 2 −π/2 4 4 2 4 16
Z r Z r−√r2 −(x−r)2
s
Z r 2 Z 0
x−r
Iy = x2 dydx = x2 r 1 − 1 − dx = (r2 )(sin θ + 1)2 r(1 − cos θ)r cos θdθ =
0 0 0 r −π/2
Z 0 Z 0
4 2 4 2 2 2 2 2 4 5π
=r (s + 2s + 1)(1 − c)cdθ = r (s c + 2sc + c − c s − 2sc − c )dθ = r 1−
−π/2 −π/2 16
Since 0 2
0
1 0 1 − cos 4θ
Z Z
1 3 1
Z Z
2 sin 2θ π
s c= s = c2 s2 = = =
−π/2 3 −π/2 3 2 4 −π/2 2 16
0 0
−1 3
Z
−1
Z
1
c = s2 = sc2 = c = −1/3
2 −π/2 2 3 −π/2
Z Z
c=1 c2 = π/4
√
Exercise 16. y = 2x, y = 0, 0 ≤ x ≤ 2, f (x − y) = |x − y|
Z 2Z x Z 2 Z √2x
2
Iy = x dydx(x − y) + x2 dydx(y − x) =
0 0 0 x
Z 2 √
Z 2
3 21 2 2 1
= x (x) − x x + x ((2x) − x2 ) − x3 ( 2x − x) =
0 2 0 2
2
1 5 2 1 4 2 1 5 2 √ 2x9/2 1 26 25 148
= x 0+ x 0− x 0− 2 + 25 = 4 − + =
10 4 10 9 0 5 9 5 45
Z 2Z x Z 2 Z √2x
Ix = y 2 dydx(x − y) + y 2 dydx(y − x) =
0 0 0 x
Z 2 x Z 2 √2x √2x !
1 3 1 4 4
1 4 1 3
= y x − y 0 dx + y −x y dx =
0 3 0 4 0 4 x 3 x
Z 2 Z 2
1 4 1 4 1 x
= x − x dx + (4x2 − x4 ) − (23/2 x3/2 − x3 ) =
0 3 4 0 4 3
Z 2 4 3/2 4
Z 2
1 4 x 2 x 1 4 23/2 5/2
= x + x2 − − x5/2 + = x + x2 − x =
0 12 4 3 3 0 6 3
1 5 1 23/2 2 7/2 24
(2 ) + 23 −
= 2 =
30 3 3 7 35
Exercise 17. Let S be thin plate of mass m.
Let the center of mass of thin plate S be located at the coordinate axis origin.
Let L0 , L be parallel to the x axis.
Let h be the perpendicular distance of L from L0 , with the sign of h included.
R b R φ (x)
Since m is the mass, a φ12(x) dydx = m
R b R φ (x)
Since CM is at (0, 0), y = 0 = a φ12(x) ydydx = 0
Z b Z φ2 (x) Z b Z φ2 (x)
moment of inertia about L = (y − h)2 dydx = y 2 − 2yh + h2 dy = IL0 + mh2
a φ1 (x) a φ1 (x)
Exercise 18. The perpendicular direction is given by the following:
125
(cos (α + π/2), sin (α + π/2)) = (− sin α, cos α)
Then
(x, y) · (− sin α, cos α) = −x sin α + y cos α = δ
We want to find the square of the above quantity, δ 2 .
q
Z b 1−( xa )2 x 2 3/2
!
cos2 α
2 2 2 2 2 2
p 3
(x sin α + −2xy sin α cos α + y cos α)dy = 2x sin αb 1 − (x/a) + 2 2b 1 −
3 a
q
2
−b 1−( xa)
x
Let = sin t
a
2
a2 a2 π
Z
1 − cos 4t
Z Z
2 2 2 2 sin 2t
a sin t cos tadt = a dt = dt =
2 4 2 8
Z π/2 Z π/2
1 + 2 cos 2t + cos2 2t π π a3π
cos3 ta cos tdt = a =a + =
−π/2 −π/2 4 4 8 8
a3 π 2 3π 1
=⇒ 2b sin2 α
+ cos2 αb3 a = πab(a2 sin2 α + b2 cos2 α)
8 3 8 4
With m = πab, the area of the ellipse, we get the desired answer.
RhRhp
Exercise 19. We want 0 0 x2 + y 2 dxdy.
s
Z hp p 2
h 2
x y p 1 p h h
ln (x + x2 + y 2 ) = h h2 + y 2 + y 2 ln ( + 1 +
x2 + y 2 dx = x2 + y 2 + )
0 2 2 2 0 y y
since, recall
1 + √ 2x 2
p
0 x +y 1 p p x2
(ln (x + x2 + y 2 )) = p =p and (x x2 + y 2 )0 = x2 + y 2 + p
x + x2 + y 2 x2 + y 2 x2 + y 2
√
s
Z h 2 Z 1 2 Z ∞
2 h h u= h
dy −−−→ y h p
2
−h 3 ln (u + 1 + u2 )
y ln + 1 + ln (u + 1 + u ) 2 du = h 4
du
0 y y ∞ u u 1 u
√ Z −3 0
u−3
Z −3
ln (u + 1 + u2 )
Z
u p p u 1
4
= ln (u + 1 + u2) = ln (u + 1 + u 2) − √
u −3 −3 −3 1 + u2
1
y=
u
If we make the following substitution, −1
du = 2 dy
y
Z y 4 −12 dy Z
u−3 u−4 −y 2
Z Z p Z p
y
√ = q = p = p dy = − y 1 + y 2− 1 + y 2 =
1 + u2 1+ 1
2 1 + y2 1 + y2
u
1 p
p p −1 p 1 p
1 + y 2 + (y 1 + y 2 + ln (y + 1 + y 2 )) =
= −y y 1 + y 2 + ln (y + 1 + y 2 )
2 2 2
q 2 s ∞
∞ 1 2
1 − 1+ u h3 √ √
−3
u p 1 1 1
h3
ln (u + 1 + u2 ) + + ln + 1 + = ( 2 + ln (1 + 2))
−3 1 3 2u 2 u u
6
1
1 3√ √ h √ 3 √ h √ 3 √
(h 2 + h3 ln (1 + 2)) + ( 2 + ln (1 + 2)) =
=⇒ ( 2 + ln (1 + 2))
4 12 3
Exercise 20. Let P0 = (0, h). We want
Z R Z √R2 −x2
√ (x2 + (y − h)2 )dydx
−R − R2 −x2
√
Z R2 −x2 p 1 p p
√ (x2 + (y − h)2 )dy = 2x2
R2 − x2 + (( R2 − x2 − h)3 − (− R2 − x2 − h)3 ) =
− R2 −x2 3
p
p 2 p 4 p 2 2
= 2x2 R2 − x2 + ((R2 − x2 )3/2 + 3 R2 − x2 h2 ) = x2 R2 − x2 + R + 2h2 R 2 − x2
3 3 3
126
x = R sin θ
Since
dx = R cos θdθ
R π/2 Z π/2 2
R4 π/2 1 − cos 4θ R4
Z Z Z
p
2 sin 2θ
x2 2 2
R −x = 2 2 2
R sin θR cos θdθ = R 4
dθ = dθ = π
−R −π/2 −π/2 2 4 −π/2 2 8
Z π/2 p Z π/2 Z π/2
2 1 + cos 2θ π
2
R −x = 2 cos θR cos θdθ(R) = R dθ = R2
−π/2 −π/2 −π/2 2 2
4 4 4 4
4R π 2 2 π R π πR πR
=⇒ + R + 2h2 R2 = + + h2 R2 π = + πR2 h2
3 8 3 2 6 3 2
Now ZZ
dydx = πr2
R2
So then the average of δ 2 is + h2
2
Exercise 21.
A = [0, 4] × [0, 1]
B = [2, 3] × [1, 3]
C = [2, 4] × [3, 4]
(1) A ∪ B
4(2, 12 ) + (2)( 52 , 42 ) 1
2 (16 + 10, 10) 13
= = ,1
4+2 6 6
(2) A ∪ C
4 1 6 7
4 2, 2 +2 2, 2 (8, 2) + (6, 7) 7 3
= = ,
4+2 6 3 2
(3) B ∪ C,
5 4 6 7
2 2, 2 +2 2, 2 11 11
= ,
2+2 4 4
(4) A ∪ B ∪ C
4(2, 12 ) + 4( 11 11
4 , 4 ) 19 13
= ,
8 8 8
Exercise 22.
rectangle R : area AR = 1(2) (x, y)R = (0, −1)
1 h
triangle T : areaAT =
(1)h = xT = 0
2 2
Z hZ ( hy −1)/(−2) Z h Z h
y y y y y2 1
y T AT = − 1 /2 ydxdy = −1+ −1 = y− = h2
0 0 h 0 −2 h h 0 h 6
1 2
1 6h
= h yT =
h/2 3
Condition for centroid to lie on the common edge, with the common edge located at the origin:
h √
h
2 3 + 2(−1)
0= =⇒ h = 2 3
h/2 + 2
Exercise 23.
Z h Z ( hy −1)(−r) Z h y y
y T AT = ydxdy = − 1 (−1) −yr − 1 dy =
0 ( hy −1)(r) 0 h h
1
isosceles triangle T :AT = 2rh = rh Z h 2
y
11 3 1 2
1
2 = 2(−r) − ydy = (−r) h − h = h2 r
0 h h 3 2 3
=⇒ y T = h2 r/3rh = h/3
1 2 π 4
semicircular disk D :AD = πr ; (2πy)A = 2πy r2 = πr3
2 2 3
127
Condition for centroid to lie in triangle:
−4r √
(rh) h3 + 12 πr2
r 2
3π
h + −2r2 ≥ 0 =⇒ h > 2r
≥0 or
(rh) + 12 πr2 3
11.22 G REEN ’ S THEOREM IN THE PLANE . S OME APPLICATIONS OF G REEN ’ S THEOREM . A NECESSARY AND
SUFFICIENT CONDITION FOR A TWO - DIMENSIONAL VECTOR FIELD TO BE A GRADIENT.
RR
Exercise 1. Green’s theorem: C P dx + Qdy = R ∂Q ∂P
R
∂x − ∂y dxdy. Thus
I ZZ
y 2 dx + xdy = (1 − 2y)dxdy
C R
R2R2
(1) 0 0 (1 − 2y)dydx = 4 − 4(2) = −4
R1 R1
(2) −1 −1 (1 − 2y)dydx = 4
R 0 R 2+x R 2 R 2−x R0 R2 R0
(3) −2 −2−x (1−2y)dydx+ 0 −2+x (1−2y)dydx = −2 (2(2+x))dx+ 0 (4−2x)dx = −2 (4+2x)dx+4(2)−4 =
8
R 2 R √4−x2 R2 √ R π/2
(4) √
−2 − 4−x2
(1 − 2y)dydx = −2
2 4 − x2 dx = −π/2 8 cos2 θdθ = 4π where we used
x = 2 sin θ
dx = 2 cos θdθ
(5) α(t) = (2 cos3 t, 2 sin3 t) = 2(cos3 t, sin3 t), 0 ≤ t ≤ 2π
Exercise 2.
2
P (x, y) = xe−y RR ∂Q
∂P
H
2 1 P dx + Qdy = ∂x − ∂y dxdy
Q(x, y) = −x2 ye−y + 2
x + y2
2 −2x
Qx = −2xye−y +
(x2 + y 2 )2
2
Py = xe−y (−2y)
Z aZ a Z a a Z a
−2x 1 1 1
2 + y 2 )2
dxdy =
2 + y2
dy = 2 + y2
− 2 =0
−a −a (x −a x −a −a a a + y2
Exercise 3. nIz = C x dy − y 3 dx = R (3x2 + 3y 2 )dydx = 3Iz n = 3
H 3 RR
Exercise 4. f = (v, u), g = ((ux − uy ), vx − vy )
(f · g) = vux − vuy + uvx − uvy = (uv)x − (uv)y = Qx − Py
ZZ Z Z Z 2π
(f · g)dxdy = (uv, uv) · ds = (uv, uv) · (− sin t, cos t)dt = (− sin2 t + sin t cos tdt = −π
R 0
Exercise 5. f, g ∈ C 1 , f, g on open connected set S in the plane.
I I ZZ
f ∇g · dα =
f gx dx + f gy dy = (fx gy + f gxy ) − (fy gx + f gyx ) =
C C R
ZZ ZZ I I
= −(gx fy + gfxy ) + (gy fx + gfyx ) = (−gfy )x − (−gfx )y = −gfx dx − gfy dy = − g(∇f ) · dα
R R
Since fxy = fyx ; gxy = gyx .
Exercise 6.
H RR RR
(1) C uvdx + uvdy = (∂x (uv) − ∂y (uv))dxdy = v(∂x u − ∂y u) + u(∂x v − ∂y v)dxdy
(2)
I ZZ
1 1
(v∂x u − u∂x v)dx + (u∂y v − v∂y u)dy = (ux vy + uvxy − vx uy − vuxy ) − (vy ux + vuyx − uy vx − uvyx ) =
2 C 2
ZZ ZZ
1
= u(vxy + vyx ) − v(uxy + uyx ) = u∂yx v − v∂yx u
2
Note the formulation of normal derivatives. Note that ds refers to the arc length.
Z Z
α(t) = (X(t), Y (t))
(P dx + Qdy) = f · T ds
C C 1
n(t) = (Y 0 (t), X 0 (t)) whenever kα0 (t)k =
6 0
T ≡ unit tangent vector to C kα0 (t)k
128
So the normal derivative is defined as
∂ψ
= ∇ψ · n
∂n
Exercise 7.
Z Z Z Z
ds dx dy
P dx + Qdy = (P, Q) · dt = P +Q dt = (QY 0 + (−P )(−X 0 ))dt =
C C dt C dt dt C
(Y 0 , −X 0 ) 0
Z Z
= (Q, −P ) · kα (t)kdt = f · nds
C kα0 (t)k C
Exercise 8.
(1)
(Y 0 , −X 0 ) 0
I I I
∂g
ds = ∇ · nds = ∇g · kα (t)kdt =
C ∂n kα0 (t)k
I I C C
ZZ ZZ
= (gx Y 0 + −gy X 0 )dt = gx dy + −gy dx = (gxx − (−gyy ))dxdy = ∇2 gdxdy
C C
(2)
(Y 0 , −X 0 ) 0
I I I
f ∇g · nds = f ∇g · kα (t)kdt = f (gx dy − gy dx) =
C kα0 (t)k
ZZ ZCZ C
ZZ
= (f gx )x − (−f gy )y = fx gx + f gxx + fy gy + f gyy = (∇f · ∇g + f (∇2 g))dxdy
I ZZ
∂g
=⇒ f ds = (f ∇2 g + ∇f · ∇g)dxdy
C ∂n R
(3) Use previous part, (b), of this exercise, Exercise 8.
I ZZ
∂g
f ds = (f ∇2 g + ∇f · ∇g)dxdy I
∂n
ZZ
C R ∂g ∂f
I =⇒ f −g ds = (f ∇2 g − g∇2 f )dxdy
∂n ∂n
ZZ
∂f C R
g ds = (g∇2 f + ∇g · ∇f )dxdy
C ∂n R
Exercise 9. P (x, y)dx + Q(x, y)dy = 0.
φx = µP
µ(x, y) is an integration factor, so µP dx + µQdy = 0 leads to φ(xy) = C s.t.
φy = µQ
dY /dt
Slope of φ(x, y) = c at (x, y) is tan θ, so tan θ = dX/dt
n = (sin θ, − cos θ)
∂φ
= ∇φ · n = (∇φ) · (sin θ, − cos θ) = φx sin θ − φy cos θ = µP sin θ − µQ cos θ =
∂n
= µ(P sin θ − Q cos θ) = µ(x, y)g(x, y)
=⇒ g = P sin θ − Q cos θ
−P P
sin θ = p or p
P 2 + Q2 P 2 + Q2 p p
=⇒ g = − P 2 + Q2 or P 2 + Q2
Q Q
cos θ = p or p
P 2 + Q2 P 2 + Q2
11.25 E XERCISES - G REEN ’ S THEOREM FOR MULTIPLY CONNECTED REGIONS . T HE WINDING NUMBER .
Exercise 1.
(1) Note that
2
x + y 2 − 2x2 x2 − y 2
−x 1 (−1)(2x) −x
∂x Q = 2 2
+ 2 2
= 2 2 2 2
= 2
x +y x x +y x +y x(x + y ) (x + y 2 )2
2
y − x2
y
∂y P = ∂y 2 2
=−
x +y (x2 + y 2 )2
129
Note that ∂x Q, ∂y P is not continuous at (0, 0).
y
P (x, y) =
x = cos t x2 + y 2
So then for −x
y = sin t Q(x, y) = 2
x + y2
Z Z 2π
P dx = sin t(− sin t)dt = −π
C 0
Z Z 2π
Qdx = − cos t(cos t)dt = −π
C 0
Z
P dx + Qdy = −2π
C
R
+ sign occurs when C is in the clockwise direction, since if x = cos t, y = − sin t, then C P dx + Qdy = 2π and
all clockwise direction, piecewise smooth Jordan curves whose interior contains (0, 0) can be deformed into a circle
(by Thm.).
(2) By Thm., we can pick any piecewise smooth Jordan curve whose interior doesn’t contain (0, 0).
Recall Green’s theorem.
Z ZZ
∂Q ∂P
P dx + Qdy = − =0
C R ∂y ∂x
∂Q ∂P
Since ∂y , ∂x , is continuous everywhere in C ∪ intC, where (0, 0) ∈
/C
Exercise 2.
x = a cos t
∂(ln r) ∂(ln r) p
f= ,− x02 + y 02 = kα0 k = a
α = (x, y)
∂y ∂x y = a sin t
p 1
ln r = ln x2 + y 2 = ln (x2 + y 2 )
2
1 1 x
(ln r)x = (2x) = 2
2 x2 + y 2 x + y2
Z
−∂(ln r) ∂(ln r) dx −∂(ln r) dy
Z Z Z
dr ∂(ln r)
f · ds = f · dt = a(− sin t) + a cos t dt = + dt =
dt ∂y ∂x ∂y dt ∂x dt
−x
Z
y
= 2 2
dx + 2 dy = −2π as shown in the previous exercise, Exercise 1.
x +y x + y2
Exercise 5.
(1) I1 − I3 = 12 − 15 = −3
(2) One possible solution is this: Draw a large curve around all 3 points, for I1 + I2 + I3 = 37. Then circle around,
inside, point 1, three times, in a clockwise fashion, to obtain −3I1 = −36.
(Y 0 (t),X 0 (t))
Exercise 6. α(t) = (X(t), Y (t)) if a ≤ t ≤ b n= √
X 02 +Y 02
b
(X(t) − x0 )Y 0 (t) − (Y (t) − y0 )X 0 (t)
Z
1
W (α0 ; P0 ) = dt =
2π
a (X(t) − x0 )2 + (Y (t) − y0 )2
Z b Z b
1 (r(t) − P0 ) · n 0 1 r(t) − P0 n
= kα (t)kdt = ds
2π a kr(t) − P0 k2 2π a kr(t) − P0 k kr(t) − P0 k
I
Ik = P dx + Qdy
Ck
1 1 1
P (x, y) = −y + 2 + =
(x − 1)2 + y 2 x + y2 (x + 1)2 + y 2
1 1 1
= −y + +
k(x, y) − (1, 0)k2 k(x, y) − (0, 0)k2 k(x, y) − (−1, 0)k2
(x − 1) x x+1
Q(x, y) = + +
k(x, y) − (1, 0)k2 k(x, y) − (0, 0)k2 k(x, y) − (−1, 0)k2
130
2
1
C1 is the smallest circle,x2 + y 2 = √
2 2
C2 : x2 + y 2 = 22
2
1 I2 = 6π
2 2
(x − 1) + y = I3 = 2π
2
2
1
C3 : x2 + y 2 =
2
2
1
(x + 1)2 + y 2 =
2
H R (x,y)−(1,0) (x,y)−(0,0) (x,y)−(−1,0)
Ik = Ck
P dx + Qdy = k(x,y)−(1,0)k2 + k(x,y)−(0,0)k2 + k(x,y)−(−1,0)k2 · nds
I2 wraps around 3 holes, (1, 0), (0, 0), (−1, 0).
I3 wraps around (0, 0) clockwise and around (1, 0), (−1, 0) counterclockwise. The wrap around (0, 0) and (1, 0) cancel each
other and so the result is we wrap around (−1, 0).
=⇒ I1 = 2π
11.28 E XERCISES - C HANGE OF VARIABLES IN A DOUBLE INTEGRAL , S PECIAL CASES OF THE TRANSFORMATION
FORMULA
Exercise 1. S = {(x, y)|x2 + y 2 ≤ a2 } where a > 0.
Recall, Z Z Z Z
··· f (φ(u))|detDφ(u)|du1 , . . . , dun = · · · f (x)dx1 , . . . , dxn φ(D) = D∗
D D
Z 2π Z a
x = r cos θ
ZZ
cos θ −r sin θ
=r>0 f (x, y)dxdy = f (r cos θ, r sin θ)rdrdθ
y = r sin θ sin θ r cos θ
S 0 0
Exercise 2. S = {(x, y)|x2 + y 2 ≤ 2x}
x − 1 = r cos θ
(x − 1)2 + y 2 = 1 2π 1
ZZ Z Z
x = r cos θ + 1 f (x, y)dxdy = f (r cos θ + 1, r sin θ)rdrdθ
(1, 0) S 0 0
y = r sin θ
Exercise 3. S = {(x, y)|a2 ≤ x2 + y 2 ≤ b2 } where 0 < a < b
ZZ Z 2π Z b
f (x, y)dxdy = f (r cos θ, r sin θ)rdrdθ
S 0 a
Exercise 4. S = {(x, y)|0 ≤ y ≤ 1 − x, 0 ≤ x ≤ 1}
1
r sin θ ≤ 1 − r cos θ =⇒ r ≤ , since sin θ, cos θ ≥ 0
sin θ + cos θ
1
Z π/4 Z sin θ+cos θ
f (r cos θ, r sin θ)rdrdθ
0 0
Exercise 5. S = {(x, y)|x2 ≤ y ≤ 1, −1 ≤ x ≤ 1} Consider imaginary angular wedges dividing up the parabolic region.
Then we identify 3 regions since each region have different boundaries.
y = 1 = r sin θ or r = csc θ
x2 = r2 cos2 θ = r sin θ or r = tan θ csc θ
Observe that y = x or θ = π4 , and y = −x or θ = −π 4 , divide up the regions.
Z π/4 Z tan θ csc θ Z 3π
4
Z csc θ Z π Z tan θ csc θ
f (r cos θ, r sin θ)rdrdθ + f (r cos θ, r sin θ)rdrdθ + f (r cos θ, r sin θ)rdrdθ
π 3π
0 0 4 0 4 0
√ p
Exercise 6. Note that 2ax − x2 = a2 − (a − x)2 . Then
x − a = r cos θ
y 2 = a2 − (a − x)2
y = r sin θ
(x − a)2 + y 2 = a2
x = (a + r cos θ) = a2 + 2ar cos θ + r2 cos2 θ
2 2
131
√ !
2a 2ax−x2 π a 2 π
2a cos θa3 a4 π 3a4 π
Z Z Z Z Z
2 2 2 2 2a π
(x + y )dy dx = dθ rdr(a + 2ar cos θ + r ) = a + dθ + =
0 0 0 0 2 0 3 4 4
Exercise 7. x = a =⇒ r cos θ = a or r = a sec θ
a x π/4 a sec θ π/4
a3 sec3 θ
Z Z p Z Z Z
2
x2 + y 2 dy dx = dθ r dr =
0 0 0 0 0 3
Z Z
3 2
Now sec θ = sec θ(1 + tan θ) and
Z
sec θ = ln | sec θ + tan θ|
sec θ tan θ + ln | sec θ + tan θ
Z
Z Z Z =⇒ sec3 θ =
2
(sec θ tan2 θ) = (sec θ)0 tan θ = sec θ tan θ − sec θ sec2 θ
√ √
a3 π/4 3 a3 ( 2 + ln | 2 + 1|)
Z
=⇒ sec θdθ =
3 0 6
Exercise 8.
Since y = r sin θ = x2 = r2 cos2 θ or r = tan θ sec θ
Z π/4 Z tan θ sec θ
1
Z π/4 √
dθ rdr = dθ tan θ sec θ = 2 − 1
0 0 r 0
Exercise 9.
a Z √a2 −y2 !
π/2 a
a4 π
Z Z Z
2 2
(x + y )dx dy = r2 rdrdθ =
0 0 0 0 8
Exercise 10. After sketching a box with vertices at (0, 0), (1, 0), (0, 1), (1, 1), it’s very clear that in polar coordinates, we must
divide the region into 2 parts by the y = x line since each region have different boundaries for r.
Z 1 Z x Z π/4 Z sec θ Z π/4 Z csc θ
f (x, y)dy dx = dθ f (r cos θ, r sin θ)rdr + dθ f (r cos θ, r sin θ)rdr since
0 0 0 0 0 0
x = 1 = r cos θ
y = 1 = r sin θ
R 2 R x √3 p R π/3 R 2 sec θ
Exercise 11. 0 x
f ( x2 + y 2 )dy dx = π/4 0 f (r)rdrdθ
x = 2 = r cos θ
r = 2 sec θ
R 1 R √1−x2 R π/2 R1
Exercise 12. 0 1−x
f (x, y)dy = 0
dθ 1 f (r cos θ, r sin θ)rdr since
sin θ+cos θ
y = 1 − x = r sin θ = 1 − r cos θ
1
r=
sin θ + cos θ
R 1 R x2 R π/4 R tan θ sec θ
Exercise 13. 0 0
f (x, y)dy dx = 0 sec θ
f (r cos θ, r sin θ)rdrdθ since
y = x2 = r sin θ = r2 cos2 θ 1 = x = r cos θ
=⇒ r = tan θ sec θ r = sec θ
x+y =u
Exercise 14. Let
x−y =v
1 1 x u x 1 1 −1 u
= =⇒ =
1 −1 y v y −2 −1 1 v
1
1
−1
J = 12 2 =
−1
2 2 2
Sketch the transformation of S to obtain a rectangle with vertices (π, π), (3π, π), (π, −π), (3π, −π).
Z 3π Z π
−1 3π 1 − cos 2u 2 3 π4
−1
Z
du dvv 2 sin2 u = π = −
π −π 2 2 π 2 3 3
132
Exercise 15.
(0, 0), (2, 10), (3, 17), (1, 7) =⇒ (0, 0), (4, 2)
4 a2 + b10
=
u = ax + by 0 c2 + d10 b = −1 a =7
=⇒
v = cx + dy = −5
0 a + 7b d=1 c
=
2 c + 7d
u+v
7 −1 x u 1 1 1 u 2 x
= = 5u+7v =
−5 1 y v 2 5 7 v 2 y
2 4 2
5u2 + 12uv + 7v 2
Z Z Z
5 3 4 3 2 4 7 2 5 28
dudv = u 0 + u v 0 + v (4)dv = 16 + 24 + = 60
0 0 8 0 24 4 8 3 3
Rr 2
Exercise 16. If r > 0, let I(r) = −r e−u du.
(1)
Z r 2 Z r Z r Z r Z r Z r Z r Z r
−u2 −x2 −y 2 2
−y 2 2
−y 2
2
I (r) = e du = e dx e dy = dy dy e−x dx = e−x dxdy
−r −r −r −r −r −r −r −r
2 2
(2) Let C1 have radius a, C2 have radius b, C1 ⊂ R ⊂ C2 , and since e−(x +y ) > 0, ∀ x, y ∈ R, then
ZZ ZZ
2 2 2 2
e−(x +y ) dxdy < I 2 (r) < e−(x +y ) dxdy
C1 C2
(3)
! a 22
! 2π 2
!
2π a 2π
e−r e−a − 1 1 − e−a
Z Z Z
−r 2 a→∞
e drrdθ = dθ = = 2π −−−→ π
0 0 0 −2 −2 2
0 0
√
I(r) → π as r → ∞
Z r Z r Z r √
2 2 √ 2 π
I(r) = e−u du = 2 e−u du = π e−u du =
−r 0 0 2
(4)
Z ∞
Γ(s) = ts−1 e−t dt
0+
Z ∞ Z ∞ Z ∞ √
1 −1/2 −t −1 −u2 −u2 π √
Γ = t e dt = u e 2udu = 2 e du = 2 = π
2 0 0 0 2
t = u2
since
dt = 2udu
x=u+v
Exercise 17.
y = v − u2
1 1
(1) detDφ = = 1 + 2u
−2u 1
(2)
(u, 0) =⇒ (u, −u2 ) u ∈ [0, 2]
(0, v) =⇒ (v, v) v ∈ [0, 2]
1 9
(u, 2 − u) =⇒ (u + 2 − u, 2 − u − u2 ) = (2, −(u + )2 + ) u ∈ [0, 2]
2 4
(3)
ZZ Z 2 Z 2−u 2 Z 2
1 2 14
dudv(1 + 2u) = (1 + 2u)dvdu = (2u − u ) + 2u(2 − u)du =
T 0 0 2 0 0 3
Z 2 Z x Z 2 2
1 2 1 3 14
dx dy = dx(x + x2 ) = x + x =
0 −x 2 0 2 3 0 3
133
(4)
ZZ Z 2 Z 2−u Z 2 Z 2−v
−2 2 −2 1 + 2u
(x − y + 1) dxdy = (u + v − v + u + 1) (1 + 2u)dudv = dv du =
S 0 0 0 0 (u2 + u + 1)2
2 2−v Z 2
−1
Z
1
= dv = dv 1 −
0 u2 + u + 1 0 0 4 − 4v + v 2 + 2 − v + 1
Now
2 5
! 2
2
4 arctan √3 (v − 2 )
Z Z 2 Z
1 1 4/3
= 5 2 3 = 2 = √ =
7 − 5v + v 2 0 (v − 2 ) + 4 3
0 √2 (v − 5 ) + 1 2/ 3
2 0
3
√
2 3 −1 5
= arctan √ − arctan √
3 3 3
√
−1
ZZ
2 3 5
=⇒ (x − y + 1)−2 dxdy = 2 − arctan √ − arctan √
S 3 3 3
x = u2 − v 2
Exercise 18.
y = 2uv
2u −2v
(1) J(u, v) = detDφ = = 4u2 + 4v 2 = 4(u2 + v 2 )
2v 2u
(2) Note the transformation of the boundaries.
x = u2 − 1 y2 y ∈ [2, 4]
(u, 1), u ∈ [1, 2] x= −1
y = 2u 4 x ∈ [0, 3]
x = 4 − v2 y2 y ∈ [4, 12]
(2, v), v ∈ [1, 3] x=4−
y = 4v 16 x ∈ [−5, 3]
x = u2 − 9 y2 y ∈ [6, 12]
(u, 3), u ∈ [1, 2] x= −9
y = 6u 36 x ∈ [−8, −5]
2 y ∈ [2, 6]
x=1−v y2
(1, v), v ∈ [1, 3] x=1−
y = 2v 4 x ∈ [−8, 0]
(3)
x2 + y 2 = u4 − 2u2 v 2 + v 4 + 4u2 v 2 = (u2 + v 2 )2 = 1 =⇒ u2 + v 2 = 1
Circle is invariant under “hyperbolic” transformation.
Z 1 Z √1−u2 Z 1 Z √
1−u2
2 2 2 2
√ (u − v )(2uv)4(u + v )dudv = 8 du √ (u2 − v 2 )(uv)dv =
−1 − 1−u2 −1 − 1−u2
Z 1 3
u 2 1
2
=8 du ((1 − u ) − (1 − u )) − u (0) = 0
−1 2 4
Exercise 19.
R
2π R R 2π (p2 +r 2 )−p+1 π
((p2 + R2 )−p+1 − (p2 )−p+1 )
ZZ Z Z
dxdy rdr 0
dθ 2(1−p) = 1−p
I(p, r) = = dθ = 0
R (ρ2 + x2 + y 2 )p 0
2
0 (p + r )
2 p 2π dθ ln 12 +R2 = π ln (1 + R2 )
R
0 2
−π 2−2p
R → ∞, if p > 1, lim I(p, r) = p
R→∞ 1−p
Exercise 20. Let u = x+y. Note that the region in xy is a rectangle starting from P = (0, −1) and spanned by a = (1, 1), b =
(−1, 1).
x 0 1 −1 1 −1
= +s +t detDφ = =1+1=2
y −1 1 1 1 1
x = s + −t
x + y = 2s − 1
y = −1 + s + t
ZZ Z 1Z 1 Z 1 Z 1
=⇒ f (x + y)dydx = f (2s − 1)2dsdt = 2 f (2s − 1)ds = f (u)du
S 0 0 0 −1
134
√
Exercise 21. Look at what we want. We eventually want ax + by = u a2 + b2 . Then try that substitution.
Also note that we want J(x, y) = detDφ 6= 0 for all points considered. We are given that a2 + b2 6= 0. Then try to get that
as a nonzero factor for J, the Jacobian.
p
ax + by = Au A = a2 + b2
=⇒ a b x Au
cx + dy = Av =
c d y Av
x 1 d −b Au
=⇒ =
y ad − bc −c a Av
We want ad − bc 6= 0 so simply use our hypothesis: a2 + b2 6= 0. Then d = a, c = −b
a −b
x 1 ua − bv
=⇒ = detDφ = Ab A =1
a
y A ub + av A A
Let’s observe how the circular region in xy changes with uv.
a2 u2 − 2abuv + b2 v 2 b2 u2 + 2abuv + a2 v 2
x2 + y 2 = 1 = + = u2 + v 2 = 1
A2 A2
Amazing! The circle is invariant under a normalized linear transformation of nonzero determinant.
Z 1 Z √1−u2 Z 1p p
=⇒ √ f (u + c)dvdu = 2 1 − u2 f (u a2 + b2 )du
−1 − 1−u2 −1
Exercise 22. From the given problem, we obviously want to make the substitution u = yx. Consider the transformed
boundaries and the Jacobian for this transformation.
xy = 1 u=1
xy = 2 u=2
x x
1
0 1
= detDφ = =
y=x u = x2 y u/x −u/x2 1/x x
y = 4x u = 4x2
Sketch the transformed region in the xu plane, with boundaries as described above. Then clearly,
Z 2 Z √u Z 2 √ Z 2 Z 2
√
1 u 1 1
du √ f (u) dx = du ln ( u) − ln f (u) = du ln u − ln u + ln 2 f (u) = ln 2 f (u)du
1 2
u x 1 2 1 2 2 1
11.34 E XERCISES - P ROOF OF THE TRANSFORMATION FORMULA IN A SPECIAL CASE , P ROOF OF THE
TRANSFORMATION FORMULA IN THE GENERAL CASE , E XTENSIONS TO HIGHER DIMENSIONS , C HANGE OF
VARIABLES IN AN n- FOLD INTEGRAL , W ORKED EXAMPLES
Exercise 1. z = xy. Note that z = 0 implied x = 0 or y = 0.
1 x xy 1 x 1
xy 2 x5 x7
ZZZ Z Z Z Z Z Z
2 3 2 3 1 1 1
xy z dxdydz = xy z dzdydx = (xy)4 dydx = dx = =
S 0 0 0 0 0 4 0 4 7 28 13 364
Exercise 2. z = 1 − x − y. z = 0 defines a boundary, so y = 1 − x.
ZZZ Z 1 Z 1−x Z 1−x−y
(1 + x + y + z)−3 dxdydz = dx dy (1 + x + y + z)−3 dz =
S 0 0 0
Z 1 Z 1−x
1/ − 2 1/ − 2
= dx dy − =
0 0 (1 + x + y + (1 − x − y))2 (1 + x + y)−2
−1 1
−1 1 −1 5
Z
1 −1 1−x
= dx (1 − x) + (1 + x + y) 0 = (1 − 1/2) + 1/2(1) − ln 2 = − ln 2
2 0 4 2 4 2 8
Exercise 3. J = r 2 sin θ (polar coordinates). Note x ≤ 0, y ≤ 0, z ≤ 0
Z π/2 Z π/2 Z 1 Z π/2 Z π/2
1
dθ dφ r cos φ sin θr sin φ sin θr cos θr sin θdr = 2
dφ sin3 θ cos θ cos φ sin φ =
0 0 0 0 0 6
Z π/2
1 1 1
= dθ sin3 θ cos θ =
6 0 2 48
135
x = au
Exercise 4. y = bv J = abc
z = cw
Z π Z 2π Z 1 Z π Z 2π
2 2 2 2 abc 2π 4πabc
=⇒ dθ dφ (u + v + w )abcr sin θdr = dθ dφ sin θ = abc(2) =
0 0 0 0 0 5 5 5
2 2 2 2
Exercise 5. z = x + y = r . z = r
ZZZ p Z 1 Z 2π Z z
11 π
x2 + y 2 dxdydz = dz dφ r2 dr = 2π =
S 0 0 0 34 6
For exercises 6,7,8, I think you have to sketch the region and surmise the new boundaries intuitively from the sketch. I don’t
see a formula you could simply plug in to determine the new regions and boundaries.
Exercise 6.
√
ZZZ Z 2 Z 2π Z 2z Z 2
2 2 1 16π
(x + y )dxdydz = r2 (rdr)dφdz = 2π (2z)2 =
S 0 0 0 0 4 3
Exercise 7.
Exercise 10.
√
ZZZ Z 2 Z 2π Z 2z Z 2
2 2 2 1 16π
(x + y )dxdydz = r (rdr)dφdz = 2π (2z)2 =
S 0 0 0 0 4 3
Exercise 13.
π 2π a
a3 4πa3
ZZZ Z Z Z
dxdydz = dθ dφ r2 sin θ = (2π)(2) =
S 0 0 0 3 3
Exercise 14.
ZZZ Z π Z 2π Z b
4π 3
dxdydz = dθ dφ r2 sin θ = (b − a3 )
S 0 0 a 3
√
Exercise 15. Let a2 + b2 + c2 = δ s.t. δ > R.
Since the sphere S of integration is rotationally symmetric, do a rotation so that so that (a, b, c) = (0, 0, δ) it’s a lot
easier!
Tip: Take advantage of symmetries, particularly spherical symmetries, and make problems easier by choosing a
convenient rotation of the coordinate axes.
Z π Z 2π Z R Z π Z 2π Z R
dθ r2 sin θdr(x2 + y 2 + (z − δ)2 )−1/2 =
dφ dθ dφ r2 sin θdr(r2 + δ 2 − 2rδ cos θ)−1/2 =
0 0 0 0 0 0
Z 2π Z R π Z 2π Z R
(r2 + δ 2 − 2rδ cos θ)1/2 r 2 2 1/2 2 2 1/2
= dφ r = dφ (r + δ + 2rδ) − (r + δ − 2rδ)
0 0 δ
0 0 0 δ
since δ > R, then ((r − δ)2 )1/2 = |r − δ| = δ − r, so then
Z R
r 4π R2
= 2π ((r + δ) − (δ − r)) =
0 δ 3 δ
x = aρ cosm θ sinn φ
m n
Exercise 16. y = bρ sin θ sin φ
z = cρ cosn φ
m n
ac θs φ
naρcm θsn−1 φcφ −maρcm−1 θsθsn φ
detJ = bsm θsn φ nbρsm θsn−1 φcφ mbρsm−1 θsn φcθ =
ccn φ −ncρcn−1 φsφ 0
= acm θsn φ(mnbcρ2 cn−1 φsφsm−1 θsn φcθ + −naρ(cm θsn−1 φcφ)(−mbcρsm−1 θsn φcn φcθ)+
+ − maρcm−1 θsθsm φ((−bncρsm θcn−1 φsn φsφ ) − nbρccn φsm θsn−1 φcφ) =
= abcmnρ2 (cm+1 (θ)sm−1 θs2n+1 φcn−1 φ + cm+1 θsm−1 θs2n−1 φcn+1 φ+
+cm−1 θsθsn φ(sm θsn+1 φcn−1 φ + sm θsn−1 φcn+1 φ)) =
= abcmnρ2 (cm+1 θsm−1 θs2n−1 φcn−1 φ + cm−1 θsm+1 θs2n−1 φcn−1 φ) =
= abcmnρ2 (cm−1 θsm−1 θs2n−1 φcn−1 φ)
136
Exercise 17.
ZZZ ZZZ ZZZ
Ix = (y 2 + z 2 )f (x, y, z)dxdydz = y 2 f (x, y, z)dxdydz + z 2 f (x, y, z)dxdydz = Ixy + Ixz
Z Z ZS Z Z ZS Z Z ZS
2 2 2
Iy = (x + z )f (x, y, z)dxdydz = x f (x, y, z)dxdydz + z 2 f (x, y, z)dxdydz = Iyz + Iyx
S S S
ZZZ ZZZ ZZZ
2 2 2
Iz = (x + y )f (x, y, z)dxdydz = x f (x, y, z)dxdydz + y 2 f (x, y, z)dxdydz = Izy + Izx
S S S
Exercise 18. The condition for the paraboloid and sphere to meet is the following:
x2 + y 2 = 4z = 5 − z 2 =⇒ z 2 + 4z − 5 = 0 or (z + 5)(z − 1) = 0
√
2 2π 5−r 2 2
r2
Z Z Z Z p
V = rdzdφdr = 2πr 5− r2 − =
0 0 r2
4 0 4
2
−1 2 3/2 4 −1 2π 3/2
(1 − 53/2 ) − 16/16) =
= 2π (5 − r ) − r /16 = 2π( (5 − 4)
3 0 3 3
Exercise 20.
2π π b
4π(b5 − a5 )
Z Z Z
1 5
r2 (r2 sin θ)drdθdφ = (b − a5 )(2)(2π) =
0 0 a 5 5
Exercise 21.
Z h Z 2π Z z Z h
2π 2 π
z dz = h3 = M
rdrdφdz =
0 0 0 0 2 3
Z h Z 2π Z z Z h
1 π
zM = rzdrdφdz = dz2π z 3 = h4
0 0 0 0 2 4
3h
z= 4so centroid is h4 away from base.
Exercise 22. Note the symmetry in φ and r.
h 2π z h
h 3 h4 πh4
Z Z Z Z
1 2 4 1
M= (h − z)rdrdφdz = dz(2π)(h − z) z = (2π) h − = (2πh ) =
0 0 0 0 2 6 8 24 12
Z h Z 2π Z z Z h 4 h
1 z 1
dz(h − z)z z 2 = π h − z 5 =
zM = (h − z)zrdrdφdz = 2π
0 0 0 0 2 4 5 0
= π(h5 )(1/4 − 1/5)
3
=⇒ z = h
5
Center of mass is 25 h from the base.
Exercise 23. Note symmetry in φ and r.
h z
2π z h
r3 2π h4 πh4
Z Z Z Z
M= rrdrdφdz = 2π dz = =
0 0 0 0 3 0 3 4 6 4h
h =⇒ z =
Z h Z 2π Z z Z h
1 2π 1 5 2π 5 5
zM = rzrdrdφdz = 2π z z3 = z = h
0 0 0 0 3 3 5 0 15
h
5 from base.
Exercise 24. Consider concentric hemispheres of radii a and b, where 0 < a < b.
π/2 2π b
b3 − a3
Z Z Z
M= dθ dφ r2 sin θ = (2π)
0 0 a 3 3 b4 − a 4
=⇒ z =
Z π/2 Z 2π Z b Z π/2
b4 − a4 π 8 b3 − a 3
zM = dθ dφ r2 sin θ(r cos θ) = dθ2π sin θ cos θ = (b4 − a4 )
0 0 a 0 4 4
Exercise 25. I tried cylindrical coordinates first. Didn’t help.
137
Tip: quickly switch and try another way, another set of coordinates, if one way doesn’t work.
Z 1 Z 1 Z 1
M= dz dy dx(x2 + y 2 + z 2 ) = 1
0 0 0
Z 1 Z 1 Z 1
2 2 2 1 1 1 1 1 7
xM = dz dy dxx(x + y + z ) = + + =
0 0 0 4 2 3 2 3 12
7
By label symmetry of x, y, z, x = y = z = 12
r a
Exercise 26. Note that z = h
Z h Z 2π Z ha z Z h 4
M 1 a M π a 4 5 M 3a2
Icone,z = r2 (rdr)dz = 2π z 4 dz = h = M
0 0 0 V 0 4 h V 10 h V 10
Z h Z 2π Z az/h
a2 1 πa2 h
V = dφ rdrdz = 2π 2 h3 =
0 0 0 2h 3 3
ZZZ ZZZ
M 2 M
Ix + Iy = (y + z 2 + x2 + z 2 )dxdydz = (x2 + y 2 + 2z 2 )dxdydz =
V V
Z h Z 2π Z az ! Z h !
M πa4 h M πa4 h 1 a2 4 M πa4 h 2πa2 5
h
2
= +2 dφ z rdrdz = +2 2π 2 z dz = + h =
V 10 0 0 0 V 10 0 2h V 10 5h2
= 2Ix
a2 2h2 a2 2h2
M 2 3M
=⇒ Ix = πa2 h
(πa h) + = +
2 3
10 5 2 10 5
Exercise 27. f = M/ 43 πR3 = M/V
Z π Z 2π Z R
2πM π R5
ZZZ Z
M 2 2 2 M 2 2
I= (x + y ) dxdydz = dθ dφ r (sin θ) r sin θ = sin θ(1 − cos2 θ) =
V 0 0 0 V V 0 5
2πM R5 4
2πM 5 1 2
= R 2 + (−2) = 4π 3 = M R2
5V 3 53R 3 5
Another way, which is quite clever, is the following. Consider that
2(x2 + y 2 + z 2 ) = x2 + y 2 + z 2 + x2 + x2 + y 2
Then ZZZ Z π Z 2π Z R
2 M
2 2 M 2 8πM 5
2 (x + y + z ) dxdydz = 2 dθ dφ r2 r sin θ = R = Iz + Iy + Ix
V 0 0 0 V 5V
5
8πM
R 2M R2
By spherical symmetry, Iz = Iy . So then I = 3 = 5
15 4πR
3
Exercise 28.
h 2π a
a3
Z Z Z
M= dz dφ rrdr = 2h(2π)
−h 0 0 3
h 2π a
3M a2
Z Z Z
1
Iz = dz dφ r2 rdrr = (2h)(2π) a5 =
−h 0 0 5 5
Exercise 29.
4πR3 2πR3
1 M
Vcap = = c= = mass density
3 2 3 V
2
1 π
Vcylinder =π 2=
2 2
π/2 2π R
2πR4 1 M πR4 M
Z Z Z
M
zM = dθ dφ = r cos θr2 sin θdr
=
0 0 0 V 4 2V 4 V
4
πR 3R
z= 3 =
4 2πR
3
8
Condition wanted is for center of mass of the mushroom to be at z = 0, for this particular choice of coordinates.
4
πR4 −π
c πR π
4 + c 2 (−1) 4 + 2
= = 0 =⇒ R4 = 2 or R = 21/4
cV + c π2 2πR3
3 + π
2
138
12.4 E XERCISES - PARAMETRIC REPRESENTATION OF A SURFACE , T HE FUNDAMENTAL VECTOR PRODUCT, T HE
FUNDAMENTAL VECTOR PRODUCT AS A NORMAL TO THE SURFACE
Exercise 1. Plane:
x = x0 + a1 u + b1 v
r(u, v) = (x0 + a1 u + b1 v)i + (y0 + a2 u + b2 v)j + (z0 + a3 u + b3 v)k =⇒ y = y0 + a2 u + b2 v
zz0 + a3 u + b3 v
Then get u, v in terms of x, y.
a1 b1 u x − x0
=
a2 b2 v y − y0
u 1 b2 −b1 x − x0 1 b2 (x − x0 ) − b1 (y1 − y0 )
=⇒ = =
v a1 b@ − b1 a2 −a 2 a1 y − y0 a1 b2 − b1 a2 −a2 (x − x0 ) + a1 (y − y0 )
So then
(a1 b2 − b1 a2 )(z − z0 ) = (a3 b2 − b3 a2 )(x − x0 ) + (b3 a1 − a3 b1 )(y − y0 )
∂u ru = (a1 , a2 , a3 )
∂v rv = (b1 , b2 , b3 )
e1 e2 e3
∂u r × ∂v r = a1 a2 a3 = (a2 b3 − a3 b2 , a3 b1 − a1 b3 , a1 b2 − a2 b1 )
b1 b2 b3
Exercise 2. Elliptic paraboloid: r(u, v) = au cos vi + bu sin vj + u2 k. Then
x = au cos v
2 2
y = bu sin v =⇒ x + y = z
a b
z = u2
∂u r = (a cos v, b sin v, 2u)
e1 e2 e3
b sin v 2u = (−2bu2 cos v, −2u2 a sin v, abu)
=⇒ a cos v
∂v r = (−au sin v, bu cos v, 0) −au sin v bu cos v 0
Exercise 3. Ellipsoid: r(u, v) = (a sin u cos v, b sin u sin v, c cos u).
x 2 y 2 z 2
=⇒ + + =1
a b c
Now
∂u r = (ac(u)c(v), bc(u)s(v), −cs(u))
∂v r = (−as(u)s(v), bs(u)c(v), 0)
e1 e2 e3
=⇒ ac(u)c(v) bc(u)s(v) −cs(u) = (bcs2 (u)c(v), acs2 (u)s(v), ab(c(u))s(u))
−as(u)s(v) bs(u)c(v) 0
Exercise 4. Surface of revolution: r(u, v) = (u cos v, u sin v, f (u)). x2 + y 2 = u2 so then
p
f ( x2 + y 2 ) = z
∂u r = (cos v, sin v, f 0 (u))
∂V r = (−u sin v, u cos v, 0)
e1 e2 e3
sin v f 0 (u) = (−f 0 u cos v, −f 0 u sin v, u)
∂u r × ∂v r = cos v
−u sin v u cos v 0
Exercise 5. Cylinder: y 2 + z 2 = a2
∂u r = (1, 0, 0)
r(u, v) = (u, a sin v, a cos v)
∂v r = (0, a cos v, −a sin v)
e1 e2 e3
=⇒ 1 0 0 = (0, a sin v, a cos v)
0 a cos v −a sin v
139
Exercise 6. Torus: r(u, v) = ((a + b cos u) sin v, (a + b cos u) cos v, b sin u), 0 < b < a.
p
x2 + y 2 = (a + b cos u)2 = (a + b2 − z 2 ) 2
∂u r = (−bs(u)s(v), −bs(u)c(v), bc(u))
∂v r((a + bc(u))c(v), (a + bc(u))(−s(v)), 0)
e1 e2 e3
−bs(u)s(v) −bs(u)c(v) bc(u) =
(a + bc(u))c(v) (a + bc(u))(−s(v)) 0
= (b(a + bc(u))c(u)s(v), b(a + bc(u))c(u)c(v), (a + bc(u))(bs(u)s2 (v) + bs(u)c2 (v)))
=⇒ (a + b cos (u))b(cos (u) sin (v), cos (u) cos (v), sin (u))
Exercise 7. r(u, v) = (a sin u cosh v, b cos u cosh v, c sinh v)
∂u r = (ac(u) cosh (v), −bs(u) cosh v, 0)
∂v r = (as(u) sinh (v), bc(u) sinh v, c cosh v)
e1 e2 e3
∂u r × ∂v r = ac(u) cosh (v) −bs(u) cos (v) 0 =
as(u) sinh (v) bc(u) sinh v c cosh v
= (−bc sin u cosh2 v, ac cos u cosh2 v, ab cos2 u cosh v sinh v + ab sin2 u cosh v sinh v)
1/2
sin2 u cos2 u sinh2 v
2
k∂u r × ∂v rk = abc cosh v + cosh v +
a2 b2 c2
Exercise 8. r(u, v) = (u + v, u − v, 4v 2 )
e1 e2 e3
∂u r = (1, 1, 0)
=⇒ 1 1 0 = (8v, −8v, −2)
∂v r = (1, −1, 8v) 1 −1 8v
p p
k∂u r × ∂v rk = 64v 2 + 64v 2 + 4 = 2 1 + 32v 2
Exercise 9. r(u, v) = ((u + v), u2 + v 2 , u3 + v 3 )
e1 e2 e3
∂u r = (1, 2u, 3u2 )
2u 3u2 = (6uv 2 − 6vu2 , 3u2 − 3v 2 , 2v − 2u) =
=⇒ 1
∂v r(1, 2v, 3v 2 ) 1 2v 3v 2
= (v − u)(6uv, −3(u + v), 2)
p
=⇒ |v − u| 36u2 v 2 + 9(u2 + 2uv + v 2 ) + 4
Exercise 10. r(u, v) = (u cos v, u sin v, 12 u2 sin 2v).
e1 e2 e3
∂u r = (c(v), s(v), us(2v))
=⇒ ∂u r × ∂v r = c(v) s(v) us(2v) =
∂v r = (−us(v), uc(v), u2 c(2v)) −us(v) uc(v) u2 c(2v)
= (u2 s(v)c(2v) − u2 c(v)s(2v), −u2 c(2v)c(v) − u2 s(2v)s(v), u) = (u2 s(−v), −u2 c(v), u)
p p
k∂u r × ∂v rk = u4 s2 (v) + u4 c2 (v) + u2 = u u2 + 1
140
12.6 E XERCISES - A REA OF A PARAMETRIC SURFACE
Exercise 2. S = r(T )
x2 + y 2 = a2 represents T .
x + y + z = a is S.
ex ey ez
∂x r = (1, 0, −1)
Using z = a − x − y, r = (x, y, a − x − y). Then =⇒ 1 0 −1 = (1, 1, 1)
∂y r = (0, 1, −1) 0 1 −1
√ √
k∂x r × ∂y rk = 1 + 1 + 1 = 3
ZZ √ √
3dxdy = 3πa2
T
y
Exercise 4. z 2 = 2xy x = 2, y = 1. Then 2z∂x z = 2y or ∂x z = z. Similarly ∂y z = xz .
e1 e2 e3
y −y −x y2 x2 z2 (x + y)2
, 1) =⇒ k∂x r × ∂y rk2 = 2 + 2 + 2 =
1 0 z =( ,
0 x z z z z z z2
1 z
r r
x+y x+y 1 x y
k∂x r × ∂y k = =√ √ =√ +
z 2 xy 2 y x
q py
√1 x
R RR
a(S) = 2 y + x dxdy so with
2 3/2
2 √2
√ √ 1 √
3x 32 2
Z r
x 4 2 −1/2 dy 4 2 1/2 8 2
R
dx =
√ = √ = y −
− −
→ 2y =
y y y 3 3 3
0 0
Z r
2 3/2
1 √
3y
y 2
R
dx 2
2 4 2
1/2
= √ = √ −−−→ 2x =
x x 3 x 3 0 3
0
√ √ !
1 8 2 4 2
=⇒ a(S) = √ + = 4
2 3 3
Exercise 5. r = (u cos v, u sin v, u2 )
a. x2 + y 2 = z. u is radius, v is angle in x − y plane.
b.
e1 e2 e3
∂u r = (c, s, 2u)
s 2u = (−2u2 c, −2u2 s, u)
=⇒ ∂u r × ∂v r = c
∂v r = (−us, uc, 0) −us uc 0
c. k∂u r × ∂v rk2 = 4u4 + u2
4
ZZ p
2 1 π π √
a(S) = u (1 + 4u2 )dudv = 2π (1 + 4u2 )3/2 = ((1 + 64)3/2 − 1) = (65 65 − 1)
3 8 0 6 6
n=6
Exercise 6. x2 + y 2 = z 2 .
x2 + y 2 + z 2 = 2ax =⇒ x2 − 2ax + a2 + y 2 + z 2 = a2 = (x − a)2 + y 2 + z 2 = a2
Determine where sphere and cone intersect: z 2 = ax, so then y 2 = ax − x2 = x(a − x). Since x > 0, a − x > 0, a > x
r r
p −a2 2
a2 a2 a 2
=⇒ y = ± x(a − x) = ± + ax − x + = − −x
4 4 4 2
Now 2z∂x z = 2x. Then
e1 e2 e3
−x −y x2 y2 z2
=⇒ k∂x r × ∂y rk2 =
x
1 0 z =( , , 1) + + =2
0 y z z z2 z2 z2
1 z
141
ZZ √ √ Z 1 Z √x(a−x) √ Z a
r
√ Z a/2
r
a2 a 2 a2
a(S) = 2dxdy = 2 √ dydx = 2 2 dx − −x =2 2 − x2 =
0 − x(a−x) 0 4 2 −a/2 4
2x
u=
a
s 2 a du = dx √ Z
√ Z a/2
a2 2 1 p
2x 2
=a 2 1− −−−−−−−−→ 1 − u2 du
−a/2 a 2 −1
u = sin θ
√ Z √ Z √
du = cos θdθ a2 2 π/2 a2 2 π/2 1 + cos 2θ 2πa2
−−−−−−−−−−→= cos2 θdθ = =
2 −π/2 2 −π/2 2 4
142
You might also like
- Essential Calculus Skills Practice Workbook With Full Solutions (Chris McMullen (McMullen, Chris) )Document365 pagesEssential Calculus Skills Practice Workbook With Full Solutions (Chris McMullen (McMullen, Chris) )Anonymous cr100% (4)
- Solution - Probability and Stochastics - CinlarDocument71 pagesSolution - Probability and Stochastics - CinlarAnshul DeshmukhNo ratings yet
- Logan PDEs Solution Manual PDFDocument78 pagesLogan PDEs Solution Manual PDFAustin FritzkeNo ratings yet
- Math Methods For Physicist Solutions Webber & ArfkenDocument33 pagesMath Methods For Physicist Solutions Webber & ArfkenJosh Brewer77% (13)
- Apostol Solution Ch7 Mathematical Analysis 2eDocument51 pagesApostol Solution Ch7 Mathematical Analysis 2eJuan Carlos Sanchez Flores100% (1)
- Solutions To Apostol Analysis Problems Vol 2Document170 pagesSolutions To Apostol Analysis Problems Vol 2Sabrina0% (1)
- Solution - Partial Differential Equations - McowenDocument59 pagesSolution - Partial Differential Equations - McowenAnuj YadavNo ratings yet
- Tom M Apostol - Mathematical Analysis Notes and Solutions (Only ch3-4-13-15-16) PDFDocument77 pagesTom M Apostol - Mathematical Analysis Notes and Solutions (Only ch3-4-13-15-16) PDFAnna RobbinsNo ratings yet
- Evans PDE Solution Chapter 6 Second-Order Elliptic EquationsDocument9 pagesEvans PDE Solution Chapter 6 Second-Order Elliptic EquationsYeltsin Acahuana50% (2)
- Solucionario Introducao A Fisica Estatistica Silvio SalinasDocument91 pagesSolucionario Introducao A Fisica Estatistica Silvio SalinasLucas RibeiroNo ratings yet
- Chapter 3 Solution.. KreyszigDocument78 pagesChapter 3 Solution.. KreyszigBladimir Blanco100% (1)
- Calculus On Manifolds (Spivak) - SolutionsDocument9 pagesCalculus On Manifolds (Spivak) - Solutionsgomdool17100% (1)
- Solutions Manual Apostol PDFDocument77 pagesSolutions Manual Apostol PDFjrsduran80% (10)
- MK 415 SolutionsToTomApostolCalculusVolume1Document281 pagesMK 415 SolutionsToTomApostolCalculusVolume1Anonymous BF5YexNo ratings yet
- Rudin SolutionsDocument22 pagesRudin SolutionsAbhijit Bhattacharya100% (6)
- PDE - Graduate Level Problems and SolutionsDocument396 pagesPDE - Graduate Level Problems and SolutionsBernard Sutopo100% (1)
- Solucionario GamelinDocument719 pagesSolucionario GamelinJhon Fabio ZuñigaNo ratings yet
- Khalil - Nonlinear Systems SlidesDocument628 pagesKhalil - Nonlinear Systems SlidesAndre Gomes88% (8)
- Steger Warming Flux Vector Splitting MethodDocument31 pagesSteger Warming Flux Vector Splitting Methodsanal_iitb100% (1)
- Spivak's Calculus On Manifolds: Solutions Manual: Thomas Hughes August 2017Document56 pagesSpivak's Calculus On Manifolds: Solutions Manual: Thomas Hughes August 2017diego2193No ratings yet
- Consider The Differential Equation 4xy'' + 2y' + Y.Document4 pagesConsider The Differential Equation 4xy'' + 2y' + Y.Jeremy Wihardi100% (1)
- Solucionario Munkres Solucionariopdf Continuous Function BasisDocument54 pagesSolucionario Munkres Solucionariopdf Continuous Function BasisMariaNo ratings yet
- Sol Modern Quantum Mechanics - J. J. Sakurai, Jim Napolitano PDFDocument123 pagesSol Modern Quantum Mechanics - J. J. Sakurai, Jim Napolitano PDFJoanSebastianPineda100% (2)
- Ar Ve Son Spectral SolutionsDocument43 pagesAr Ve Son Spectral Solutionssticker592100% (1)
- Arfken Weber Math Selected Problem SolutionDocument79 pagesArfken Weber Math Selected Problem Solutionalirezatehrani79% (14)
- 40 Functional Equations Problem 16 Solution - Adib HasanDocument1 page40 Functional Equations Problem 16 Solution - Adib HasanK. M. Junayed AhmedNo ratings yet
- AlgebraicDocument3 pagesAlgebraicYU CHOOI YING MoeNo ratings yet
- Worksheet Fourier SeriesDocument6 pagesWorksheet Fourier SeriesInnocent MbeleNo ratings yet
- Sec 1-1Document10 pagesSec 1-1Metin DenizNo ratings yet
- 2D Geometrical TransfDocument13 pages2D Geometrical TransfDr Sameer Chakravarthy VVSSNo ratings yet
- Polynomials J Ainsworth Dec21Document4 pagesPolynomials J Ainsworth Dec21Jason Wenxuan MIAONo ratings yet
- Mate 2Document16 pagesMate 2Trifan_DumitruNo ratings yet
- UPSC Civil Services Main 1985 - Mathematics Calculus: Sunder LalDocument5 pagesUPSC Civil Services Main 1985 - Mathematics Calculus: Sunder Lalsayhigaurav07No ratings yet
- Math MSC PS02 SolutionsDocument3 pagesMath MSC PS02 SolutionsAhmed AdelNo ratings yet
- Equations With Regular-Singular Points (Sect. 5.5) .: RecallDocument14 pagesEquations With Regular-Singular Points (Sect. 5.5) .: RecallBhat MusaibNo ratings yet
- HW140B 2 SolutionsDocument6 pagesHW140B 2 SolutionsIdris AddouNo ratings yet
- Tutorial Sheet 2Document3 pagesTutorial Sheet 2Psycho MindNo ratings yet
- Solutions Problem Set 1 (Diff)Document4 pagesSolutions Problem Set 1 (Diff)AdityaNo ratings yet
- MATH2230 Sol2Document3 pagesMATH2230 Sol2Sukmana PutraNo ratings yet
- Vectorspace PDFDocument8 pagesVectorspace PDFTatenda BizureNo ratings yet
- Notes On Discontinuous F (X) Satisfying F (X + Y) F (X) F (Y)Document5 pagesNotes On Discontinuous F (X) Satisfying F (X + Y) F (X) F (Y)cNo ratings yet
- 87 S1Document7 pages87 S1ParmenNo ratings yet
- Galois Solutions PDFDocument23 pagesGalois Solutions PDFAbir KayalNo ratings yet
- E textofChapter2Module3Document9 pagesE textofChapter2Module3Dr.S. JACKSONNo ratings yet
- Lecture 10 NotesDocument5 pagesLecture 10 Notesalex 2000No ratings yet
- The Transcendence of π: Steve Mayer November 2006Document6 pagesThe Transcendence of π: Steve Mayer November 2006Fernando CarvalhoNo ratings yet
- IGCSE Syllabus OverviewDocument26 pagesIGCSE Syllabus OverviewHaris MdNo ratings yet
- Table of DerivativesDocument5 pagesTable of DerivativesZlatan EgiptoNo ratings yet
- Berkeley Math Circle: Monthly Contest 2 SolutionsDocument3 pagesBerkeley Math Circle: Monthly Contest 2 SolutionshungkgNo ratings yet
- 5th SessionDocument5 pages5th SessionAlirezaNo ratings yet
- Expanding Subspace TheoremDocument7 pagesExpanding Subspace TheoremA kumarNo ratings yet
- Metric Space and Norm Linear Space Important QuestionsDocument3 pagesMetric Space and Norm Linear Space Important QuestionsP GOSWAMINo ratings yet
- 2060A Ex 1 2019Document4 pages2060A Ex 1 2019Samuel Alfonzo Gil BarcoNo ratings yet
- Galois Solutions PDFDocument20 pagesGalois Solutions PDFlaxmi mahtoNo ratings yet
- Cauchy Functional EquationDocument4 pagesCauchy Functional Equationaruodas100% (1)
- Week6 NotesDocument14 pagesWeek6 NotesReevu ThapaNo ratings yet
- Homework 2Document4 pagesHomework 2Apel KingNo ratings yet
- Random Variable Unit I NotesDocument8 pagesRandom Variable Unit I Notessneha srinivasanNo ratings yet
- Sheet01 H LSGDocument6 pagesSheet01 H LSGAusten Meusel100% (1)
- Formula RioDocument2 pagesFormula Rioricardorita2k5No ratings yet
- Acharya Institute of Graduate Studies: Questions CarryingDocument4 pagesAcharya Institute of Graduate Studies: Questions CarryingRegina SzlúkaNo ratings yet
- Motion in Straight Line 3 Physics Class NotesDocument15 pagesMotion in Straight Line 3 Physics Class NotesAtul SanghaviNo ratings yet
- Aersys 7024Document3 pagesAersys 7024antonellodelreNo ratings yet
- Introduction, Linear EquationDocument9 pagesIntroduction, Linear Equationshahmeerk736No ratings yet
- Instant Download Mathematical quantization 1st Edition Nik Weaver PDF All ChaptersDocument61 pagesInstant Download Mathematical quantization 1st Edition Nik Weaver PDF All Chaptersshahloaldiqs100% (3)
- 2022 Tutorial TestsDocument7 pages2022 Tutorial TestsSakhile GumedeNo ratings yet
- Mat3100 Lecture Notes 6 2018 PDFDocument7 pagesMat3100 Lecture Notes 6 2018 PDFKelvin BwalyaNo ratings yet
- JacobiansDocument4 pagesJacobiansAnoushka BansalNo ratings yet
- The Dirac Equation and Gamma MatricesDocument3 pagesThe Dirac Equation and Gamma MatricesPreeti TejaNo ratings yet
- Notes On "Integration Theory: A Second Course" by Martin VathDocument4 pagesNotes On "Integration Theory: A Second Course" by Martin Vathdconn_1No ratings yet
- Solutions of Exercises In: Dongkwan KimDocument31 pagesSolutions of Exercises In: Dongkwan Kimalgohrythm100% (1)
- Unit 5 Transformation NotesDocument33 pagesUnit 5 Transformation NotesTarun BambhaniyaNo ratings yet
- 3-Vehicle Equation of MotionDocument16 pages3-Vehicle Equation of MotionBira-aanuu SamuelNo ratings yet
- AdilaDocument15 pagesAdilamr FadhilNo ratings yet
- Numerical AnalysisreportDocument11 pagesNumerical AnalysisreportTas IR Ow AisNo ratings yet
- CAD & A Second Series QPDocument1 pageCAD & A Second Series QPimnithinNo ratings yet
- Mat 1320 Exercises-Solutions 4-5Document4 pagesMat 1320 Exercises-Solutions 4-5eyyupkaya2003No ratings yet
- Large StrainDocument28 pagesLarge StrainMoumita SitNo ratings yet
- Pages From (S. S. Quek, G.R. Liu) The Finite Element Method (BookZZ - Org) 2-2Document3 pagesPages From (S. S. Quek, G.R. Liu) The Finite Element Method (BookZZ - Org) 2-2mehdiqNo ratings yet
- Information Processing Letters: Malik Magdon-IsmailDocument4 pagesInformation Processing Letters: Malik Magdon-IsmailNo12n533No ratings yet
- MSG.17.Matrices and Matrix TransformationsDocument14 pagesMSG.17.Matrices and Matrix TransformationsDahanyakage WickramathungaNo ratings yet
- Inn 1 - 21ma44t Nm-ChemicalDocument10 pagesInn 1 - 21ma44t Nm-Chemicalsabarisbsabari07No ratings yet
- "Thinking Quantum": Lectures On Quantum Theory For High-School StudentsDocument38 pages"Thinking Quantum": Lectures On Quantum Theory For High-School StudentsJhonathan BarriosNo ratings yet
- Bachelor of Computer Science (Hons) 2013-14-1st YearDocument10 pagesBachelor of Computer Science (Hons) 2013-14-1st YearBalayet Hossain PulakNo ratings yet
- 1 VectorsDocument56 pages1 VectorsSam Tries StuffNo ratings yet
- Advanced Engineering Mathematics SyllabusDocument7 pagesAdvanced Engineering Mathematics SyllabusDr. Samer As'adNo ratings yet
- 4-UNIT-Matrix Inversion-Iterative MethodsDocument11 pages4-UNIT-Matrix Inversion-Iterative MethodsELEFTHERIOS GIOVANISNo ratings yet
- Solutions of Practice Questions Matrices and DeterminantDocument4 pagesSolutions of Practice Questions Matrices and DeterminantSyed Abdul Mussaver ShahNo ratings yet
- Question Bank For Module 3Document3 pagesQuestion Bank For Module 3Raj GaneshNo ratings yet