11/1/2017 Second Hourly Practice 8 Math 21a, Fall 2017

Name:

• Start by printing your name in the above box

MWF 9 Jameel Al-Aidroos and check your section in the box to the
left.
MWF 9 Dennis Tseng
• Do not detach pages from this exam packet
MWF 10 Yu-Wei Fan
or unstaple the packet.
MWF 10 Koji Shimizu
• Please write neatly. Answers which are illeg-
MWF 11 Oliver Knill ible for the grader cannot be given credit.
MWF 11 Chenglong Yu • Show your work. Except for problems 1-
MWF 12 Stepan Paul 3,8, we need to see details of your computa-
tion.
TTH 10 Matt Demers
• All functions can be differentiated arbitrarily
TTH 10 Jun-Hou Fung
often unless otherwise specified.
TTH 10 Peter Smillie
• No notes, books, calculators, computers, or
TTH 11:30 Aukosh Jagannath other electronic aids can be allowed.
TTH 11:30 Sebastian Vasey • You have 90 minutes time to complete your
work.

1 20

2 10

3 10

4 10

5 10

6 10

7 10

8 10

9 10

10 10

Total: 110
Problem 1) True/False questions (20 points), no justifications needed

Every function f (x, y) of two variables has either a global minimum or a

1) T F global maximum.

Solution:
Take for example f (x, y) = x + y. This function has a constant nonzero gradient and so
no critical point. It is unbounded above and below.

The linearization of the function f (x, y) = ex+3y at (0, 0) is L(x, y) = 1 +

2) T F x + 3y.

Solution:
Use the definition of linearization. The gradient of f is ∇f = hex+3y , 3ex+3y i. At (0, 0)this
is h1, 3i. We have f (0, 0) = 1 so that L(x, y) = 1 + x + 3y.

The function f (x, y, z) = x2 cos(z) + x3 y 2 z + (y − 2)3 y 5 satisfies the partial

3) T F differential equation fxyxzxy = 12.

Solution:
Use Clairot.

4) T F If xez = y 2 z, then ∂z/∂x = ez /(y 2 − xez ).

Solution:
This is a direct application of implicit differentiation zx = −fx /fz .

5) T F The function cos(x2 ) cos(y 2 ) has a local maximum at (0, 0).

Solution:
The value at (0, 0) is equal to 1. The functions and so the product take values between
−1 and 1.

R π/4 R 2
The value of the double integral 0 0 x3 cos(y) dxdy is the same as
6) T F R π/4
( 02 x3 dx)( 0 cos(y) dy).
R
 Solution:
The function cos(y) is a constant for the inner integral so that we can pull it out of the
inner integral.

7) T F The gradient of f (x, y) is always tangent to the level curves of f .

Solution:
It is perpendicular

If f (x, y, z) = x − 2y + z, then
√ the largest possible directional derivative
8) T F D~u f at any point in space is 6.

Solution: √
The gradient has length 6. The directional derivative into the direction of the gradient
is the length of the gradient.

R1 R1 R1 R1
9) T F 0 0 (x2 + y 2) dxdy = 0 0 r 3 drdθ.

Solution:
While the substitution of the function and the r factor have been done correctly, the
region changes. The right integral defines a sector, while the left integral is an integral
over the unit square.

It is possible that the directional derivative D~v f is positive for all unit
10) T F vectors ~v .

Solution:
The directional derivative changes sign if ~v is replaced by −~v .

Using linearization of f (x, y) = xy we can estimate f (0.999, 1.01) ∼ 1 −

11) T F 0.001 + 0.01 = 1.009.

Solution:
L(x, y) = 1 − 1 · 0.001 + 1 · 0.01.
 d
12) T F Given a curve ~r(t) on a surface g(x, y, z) = −1, then dt
g(~r(t)) < 0.

Solution:
It is zero.

13) T F If f (x, y) has a local minimum at (0, 0) then it is possible that fxy (0, 0) > 0.

Solution:
2
D = fxx fyy − fxy > 0 is still possible, if fxx and fyy are large. For example x2 + y 2 + xy/10
has a local minimum at (0, 0) even so fxy > 0.

14) T F The function f (x, y) = −x8 − 2x6 − y 8 has a local minimum at (0, 0).

Solution:
One can not use the second derivative test because the discriminant is zero. But the
function is zero at (0, 0) and strictly negative everywhere else. Therefore, (0, 0) is a global
maximum. It is definitly not a minimum.

If ~r(t) is a curve in space and f is a function of three variables, then

15) T F d
f (~r(t)) = 0 for t = 0 implies that ~r(0) is a critical point of f (x, y, z).
dt

Solution:
We can have r(t) = (t, 0, 0) and f (x, y, z) = x2 + (y − 1)2 .

Let a, b, c be the number of saddle points, maxima and minima of a function

16) T F f (x, y). Then a ≤ b + c.

Solution:
Already x2 − y 2 is a counter example.

If f (x, y) is a nonzero function of two variables and R is a region, then

R f (x, y) dxdy is the volume under the graph of f and therefore a positive
RR
17) T F
value.

Solution:
if f is replaced by −f , then the sign of the integral changes too.
 We extremize f (x, y) under the constraint g(x, y) = c and obtain a solution
18) T F (x0 , y0). If the Lagrange multiplier λ is positive, then the solution is a
minimum.

Solution:
There is no relation between the sign of λ and minima and maxima. Change g = c to
−g = −c and the sign of λ changes.

The tangent plane to a surface f (x, y, z) = 1 intersects the surface in exactly

19) T F one point.

Solution:
Take a one sheeted hyperboloid.

Let ~v be a vector of length 1 in space. Given a function f (x, y, z) of three

20) T F variables. If (x0 , y0 , z0 ) is a critical point of f , then it is a critical point of
g(x, y, z) = Dv f (x, y, z).

Solution:
Let ~v = h1, 0, 0i. Now g(x, y, z) = fx (x, y, z) and ∇g = hfxx , fxy , fxz i.

Problem 2) (10 points)

a) (6 points) Match the regions with the corresponding double integrals

1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

a 0.0 0.2 0.4 0.6 0.8 1.0 b 0.0 0.2 0.4 0.6 0.8 1.0 c 0.0 0.2 0.4 0.6 0.8 1.0
 1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

d 0.0 0.2 0.4 0.6 0.8 1.0 e 0.0 0.2 0.4 0.6 0.8 1.0 f 0.0 0.2 0.4 0.6 0.8 1.0

Enter a,b,c,d,e or f Integral of f (x, y) Enter a,b,c,d,e or f Integral of f (x, y)

R 1 R √x R 1 R √1−x2
0 x2 f (x, y) dydx 0 0 f (x, y) dydx

R 1 R √y R1 R1
0 0 f (x, y) dxdy 0 (1−x)2 f (x, y) dydx

R1 R1 R 1 R √1−x2
0 y2 f (x, y) dxdy 0 (1−x)2 f (x, y) dydx

b) (4 points) Match the PDE’s with the names. No justifications are needed.
Enter A,B,C,D here PDE Enter A,B,C,D here PDE
fxx = −fyy fxx = fyy

fx = fy fx = fyy

A) Wave equation B) Heat equation C) Transport equation D) Laplace equation

Solution:
a b
a) c f
d e
D A
b)
C B

Problem 3) (10 points)

a) (3 points) Find and classify all the critical points of f (x, y) = xy − x on the plane.

b) (2 points) Decide whether an absolute maximum or an absolute minimum of f exists

on the plane R2 .

c) (3 points) Use the method of Lagrange multipliers to find the maximum and minimum
of f on the boundary x2 + 4y 2 = 12 of the elliptical region G : x2 + 4y 2 ≤ 12.

d) (2 points) Find the absolute maximum and absolute minimum of f on the region G
given in c).

Solution:
a) ∇f = hy − 1, xi = ~0 for (x, y) = (0, 1). Since fxx = fyy = 0 and fxy = 1, the
discriminant is D = 02 − 22 < 0 and (0, 1) is a saddle point .
b) There is no global maximum, nor any global minimum on the plane. On the x-axes
y = 0 for example, we have f (x, 0) = −x which is unbounded both from above and from
below.
c) The Lagrange equations are

y − 1 = λ 2x
x = λ 8y
x + 4y 2 = 12 .
2

y 6= 0, because otherwise the second equation would give x = 0, contradicting the con-
straint. Also x 6= 0, because otherwise, the first equation would give y = 1, again contra-
dicting the constraint. Dividing the first by the second gives (y−1)/x = (1/4)x/y or 4y(y−
1) = x2 . Plugging this into the constraint gives 4y(y − 1) + 4y 2 = 12. The solutions of this
√ √
quadratic equation are y = 3/2 or y = −1. The extrema are (±2 2, −1) and (± 3, 1.5) .
√ √ √ √ √ √ √
Since f (2 2, −1) = −4 2, f (−2 2, −1) = 4 2, f ( 3, 1.5) = 23 and f (− 3, 1.5) =
√ √ √
− 23 , the maximum is (x, y) = (−2 2, −1) and the minimum is (x, y) = (2 2, −1) .
d) From parts (a) and (c) we have a list of all candidates for global extrema. The global
√ √
maximum value of f on G is f (−2 2, −1) = 4 2 , the global minimal value on G is
√ √
f (2 2, −1) = −4 2 .

Problem 4) (10 points)

Find the cylindrical basket which is open on the top has has the largest volume for fixed
area π. If x is the radius and y is the height, we have to extremize f (x, y) = πx2 y under
the constraint g(x, y) = 2πxy + πx2 = π. Use the method of Lagrange multipliers.
Solution:
The Lagrange equations are
2xyπ = (2xπ + 2yπ)λ
πx2 = 2πxλ
πx2 + 2πxy = π

Since x = 0 is not possible (it would violate the constraint), we can divide the second
√
equations by x and divide the first by the second equation. This gives x = y = 1/ 3 .
√
The maximum value is π 3/9 .

Problem 5) (10 points)

The Pac-Man region R is bounded by the lines y = x, y = −x and the unit circle. The
number
x dxdy
RR
a = R RR
R 1 dxdy
defines the point C = (a, 0) called center of mass of the region. Find it.
Solution:
Z 7π/4 Z 1
7π/4
√
r cos(θ) rdrdθ = (1/3) sin(θ)|π/4 = − 2/3 .
π/4 0
Z 7π/4 Z 1
rdrdθ = (1/2)(7π/4 − π/4) = 6π/8 = 3π/4 .
π/4 0

The second integral is the area of the Pac-Man, which is 3/4 of the area of the full disc.
√
Dividing the first by the second integral gives the result a = −4 2/(9π) . The center of
√
mass is (−4 2/(9π), 0) .

Problem 6) (10 points)

√
a) (5 points) Find the tangent plane to the surface xyz = 60 at (x, y, z) = (100, 36, 1).
√
b) (5 points) Estimate 100.1 ∗ 36.1 ∗ 0.999 using linear approximation. Here, for clarity
reasons, we use ∗ for the usual multiplication for numbers.

Solution:
a) We have s
yz xz xy
r r
∇f (x, y, z) = h , , i/2
x y z
6 10
∇f (100, 36, 1) = h , , 60i/2
10 6
The tangent plane is (3/10)x + (5/6)y + 30z = 90 . We have obtained the constant on
the right by plugging in the point (x, y, z) = (100, 36, 1).
b) Since f (100, 36, 1) = 60, we have L(x, y, z) = 60+(3/10)(x−100)+(5/6)(y−36)+30(z−
1). We have L(100.1, 36.1, 0.999) = 60+0.03+0.08333..−0.03 = 60.08333... = 60 + 1/12 .
This is very close to the actual value 60.0832455.... You have in this problem computed
the square root of a real number by hand with an accuracy of 4 digits after the comma.

Problem 7) (10 points)

Oliver got a diagmagnetic kit, where strong magnets produce a force field in which pyrolytic
graphic flots. The gravitational field produces a well of the form f (x, y) = x4 +y 3 −2x2 −3y.
Find all critical points of this function and classify them. Is there a global minimum?
 Right picture credit: Wikipedia.

Solution:
To find the critical points, we have to solve the system of equations fx = 4x3 − 4x =
0, fy = 3y 2 − 3 = 0. The first equation gives x = 0 or x = ±1. The second equation
fy = 3y 2 − 3 = 0 gives y = ±1. There are 3 · 2 = 6 critical points. We compute the
discriminant D = 6y(12x2 − 4) and fxx = 12x2 − 4 at each of the 6 points and use the
second derivative test to determine the nature of the critical point.
point D fx x nature value
(-1, -1) -48 8 saddle 1
(-1, 1) 48 8 min -3
( 0, -1) 24 -4 max 2
( 0, 1) -24 -4 saddle -2
( 1, -1) -48 8 saddle 1
( 1, 1) 48 8 min -3

There is no global minimum, nor any global maximum since for x = 0, the function is
f (0, y) = y 3 − 3y which is unbounded from above and from below (it goes to ±∞ for
y → ±∞).

Problem 8) (10 points)

Let f (x, y) = xy.

a) (2 points) Find the direction of maximal increase at the point (1, 1).

b) (3 points) Find the directional derivative at (1, 1) in the direction h3/5, 4/5i.
√ √
c) (2 points) The curve ~r(t) = h 2 sin(t), 2 cos(t)i passes through the point (1, 1) at
some time t0 . Find dtd f (~r(t)) at time t0 directly.

d
d) (3 points) Find dt
f (~r(t)) at time t0 using the multivariable chain rule.
Solution: √
a) ∇f (x, y) = hy, xi, ∇f (1, 1) = h1, 1i. The direction of maximal increase is h1, 1i/ 2 .
b) Dv f (1, 1) = h1, 1i · h3/5, 4/5i = 7/5 .
c) It is at the time t0 = π/4, where the curve passes through thte point (1, 1). We have
f (~r(t)) = 2 cos(t) sin(t) = sin(2t) and d/dtf (~r(t)) = 2 cos(2t) which is 0 at time t = π/4.
d) By the multi variable chain rule, dtd f (~r(t)) = ∇f (1, 1) · h− sin(π/4), cos(π/4)i = 0 .

Problem 9) (10 points)

Integrate the function

y5 − 1
f (x, y) = 1/3
y − y 1/4
on the finite region bounded by the curves y = x3 and y = x4 .

Solution:
Make a picture! The two graphs intersect at 0 and 1 forming a grass shaped region.

The type I integral

1 x3 y5 − 1
Z Z
dydx
0 x4 y 1/3 − y 1/4
can not be evaluated (at least not without going through difficult substitution/partial
fraction procedures which can fill pages).
We decide therefore, to change the order or integration and write a type II integral:
Z 1 Z y 1/4 y5 − 1
dxdy
0 y 1/3 y 1/3 − y 1/4
R1
Now the inner integral can be solved and give (1 − y 5). We end up with 0 (1 − y 5 ) dy =
5/6 .
Problem 10) (10 points)

The main building of a mill has a cone shaped roof and

cylindrical walls. If the cylinder has radius r, the height
of the side wall is h and the height of the roof is h, then
the volume is
V (h, r) = πr 2 h + hπr 2 /3 = (4π/3)hr 2
and assume the cost of the building is
A(h, r) = πr 2 + 2πrh + π2r 2 = π(3r 2 + 2rh)

which is the area of the ground plus the area of the

wall plus 2πrh, the cost for the roof. For fixed volume
V (h, r) = 4π/3, minimize the cost A(h, r) using the La-
grange multiplier method.

Solution:
After dividing out some constants and taking g = hr 2 = 1, the Lagrange equations become
6r + 2h = λ2hr
2r = λr 2
r2h = 1
The second equation can be divided by r since r = 0 is incompatible with the third
equation. The first can be divided by 2. We get
3 ∗ r + h = λhr
2 = λr
2
r h = 1
You can plug in λr from the second equation into the first to get
3r + h = 2h
r2h = 1

The first equation shows h = 3r and plugging this into the third equation gives r = 1/31/3
and h = 3r = 32/3 .