Department of Mathematics

Indian Institute of Technology, Bombay

Powai, Mumbai400 076, INDIA.

MA 203Mathematics III
Autumn 2006

Instructors:
A.
N.
R.
G.

Name

Athavale
Nataraj
Raghunathan (Instructor in-charge)
K. Srinivasan

Roll No :

Course contents of MA 203 (Mathematics III):

Ordinary differential equations of the 1st order, exactness and integrating factors, variation of parameters, Picards iteration method. Ordinary linear differential equations of the nth order, solution of
homogeneous and non-homogeneous equations. The operator method. The methods ofundetermined coefficients and variation of parameters. Systems of differential equations, Phase plane. Critical points,
stability.
Infinite sequences and series of real and complex numbers. Improper integrals. The Cauchy criterion,
tests of convergence, absolute and conditional convergence. Series of functions. Improper integrals depending on a parameter. Uniform convergence. Power series, radius of convergence. Power series methods for
solutions of ordinary differential equations. Legendre equation and Legendre polynomials, Bessel equations
and Bessel functions of first and second kind. Orthogonal sets of functions. Sturm-Liouville problems. Orthogonality of Bessel functions and Legendre polynomials. The Laplace transform. The Inverse transform.
Shifting properties, convolutions, partial fractions. Fourier series, half-range expansions. Approximation
by trignometric polynomials. Fourier integrals.

Texts/References
E. Kreyszig, Advanced Engineering Mathematics, 8th ed., Wiley Eastern, 1999.

Teaching Plan
[K] refers to the text book by E. Kreyszig, Advanced Engineering Mathematics, 8th Edition, John Wiley
and Sons(1999).

Policy for Attendance

Attendance in both lectures and tutorial classes is compulsory. Students who fail to attend 80% of the
lectures and tutorial classes may be awarded an XX grade.

Evaluation:

Figures in parentheses denote the percentage of the marks assigned to each quiz or exam.

Quiz 1 (16.67): 6:00 - 7:00

Quiz 2 (16.67): 6:00 - 7:00
Mid-semester examination
End-semester examination

p.m., Wednesday, August 16.

p.m., Wednesday, October 18
(33.33)
(33.33)

Topics to be covered before the mid-semester examination

No.
1.

2.

3.

4.

Topic
Basic concepts and ideas,
1st order linear equations (homogeneous
and non homogeneous),
Separation of Variables,
Exact equations, integrating factors
The Bernoulli Equation
Existence and Uniqueness: Picards iteration
Singular solutions, enveloppes and
orthogonal trajectories
Linear Differential equations:
Homogeneous equations with constant
coefficients,
Existence and Uniqueness, The Wronskian,
Non-homogeneous equations

in [K]
1.1-1.2

No.of Lec.

1.3 - 1.4

1.5-1.7
1.9
4
1.8

2.1-2.3

2.7-2.8

5.

The Cauchy-Euler equations

The Method of undetermined coefficients,
The method of variation of parameters
Sequences and Series, Convergence Tests

2.92.10
2.11-2.12
14.1, A- 3.3

4
2
3

6.

Uniform convergence, and Power Series

14.2 - 14.5

Topics to be covered after the mid-semester examination

No.

Topic

in [K]

No.of Lec.

7.

Improper integrals (p. 27 of this booklet)

8.
9.

Laplace transforms and systems of equations

Series solutions: Legendres equation
and the Legendre polynomials

5.1-5.7

4.3

10.

Frobenius method, the Bessels functions

4.44.6

10.

Sturm-Liouville Problems: Eigenvalues

and eigenfunctions properties and more

4.7-4.8

3
4

11.

Fourier Series and Fourier Integrals

10.1 - 10.10

12.

Qual. Theory: Phase plane analysis,

Critical Points, Linearized Stability.

3.3 - 3.5

Tutorial Sheet No. 1

Q.1. Classify the following equations (order, linear or non-linear):

d3 y
dy
d2 y
dy 2
dy
(i) dx
3 + 4( dx ) = y (ii) dx + 2y = sin x (iii) y dx2 + 2x dx + y = 0
(iv)

d4 y
dx4

dy
+ x2 y = 0. (v)(1 + y 2 ) ddt2y + t ddt6y + y = et .
+ (sin x) dx

Q.2. Formulate the differential equations represented by the following functions by eliminating the arbitrary constants a, b and c:
(i) y = ax2 (ii) y a2 = a(x b)2 (iii) x2 + y 2 = a2 (iv) (x a)2 + (y b)2 = a2
(v) y = a sin x + b cos x + a (vi) y = a(1 x2 ) + bx + cx3 (vii) y = cx + f (c).
Also state the order of the equations obtained.

Q.3. Solve the equation x3 (sin y)y 0 = 2. Find the particular solution such that y(x)

as x +.

Q.4. Prove that a curve with the property that all its normals pass through a point is a circle.
Q.5. Find the values of m for which
(a) y = emx is a solution of
(i) y 00 + y 0 6y = 0 (ii) y 000 3y 00 + 2y 0 = 0.

(b) y = xm for x > 0 is a solution of

(i) x2 y 00 4xy 0 + 4y = 0 (ii) x2 y 000 xy 00 + y 0 = 0.
Q.6. For each of the following linear differential equations verify that the function given in brackets is a
solution of the differential equation.
(i) y 00 + 4y = 5ex + 3 sin x (y = a sin 2x + b cos 2x + ex + sin x)
(ii) y 00 5y 0 + 6y = 0, (y1 = e3x , y2 = e2x , c1 y1 + c2 y2 )

(iii) y 000 + 6y 00 + 11y 0 + 6y = e2x (y = aex + be2x + ce3x xe2x )

(iv) y 000 + 8y = 9ex + 65 cos x, (y = ae2x + ex (b cos 3x + c sin 3x) + 8 cos x sin x + ex )
Q.7. Let i be a solution of y 0 + ay = bi (x) for i = 1, 2.
Show that 1 + 2 satisfies y 0 + ay = b1 (x) + b2 (x).
y 0 + y = sin x + 3 cos 2x passing through the origin.

Use this result to find the solutions of

Q.8. Obtain the solution of the following differential equations:

(i) (x2 + 1)dy + (y 2 + 4)dx = 0; y(1) = 0

(ii) y 0 = y cot x; y(/2) = 1

(iii) y 0 = y(y 2 1), with y(0) = 2 or y(0) = 1, or y(0) = 0

yx
= 0; y(1) = 1
(v) y 0 +
y+x

(iv) (x + 2)y 0 xy = 0; y(0) = 1

(vii) 2(y sin 2x + cos 2x)dx = cos 2xdy; y() = 0.

(vi) y 0 = (y x)2 ; y(0) = 2

(viii) y 0 =

1
(x + 1)(x2 + 1)

Q.9. For each of the following differential equations, find the general solution (by substituting y = vx)
(i) y 0 =

y 2 xy
x2 + xy

(iii) xy = y + x cos2 (y/x)

(ii) x2 y 0 = y 2 + xy + x2
(iv) xy 0 = y(ln y ln x)

Q.10. Show that the differential equation

reduced to

ax + by + m
dy
=
where a, b, c, d, m and n are constants can be
dx
cx + dy + n

dy
ax + by
=
if ad bc 6= 0. Then find the general solution of
dt
cx + dy

(i) (1 + x 2y) + y 0 (4x 3y 6) = 0

(ii) y 0 =

yx+1
yx+5

(iii) (x + 2y + 3) + (2x + 4y 1)y 0 = 0.

1
Q.11. Solve the differential equation 1 y 2 dx+ 1 x2 dy = 0 with the conditions y(0) =
3. Sketch
2
the graphs of the solutions and show that they are each arcs of the same ellipse. Also show that after
these arcs are removed, the remaining part of the ellipse does not satisfy the differential equation.
Q.12. The differential equation y = xy 0 + f (y 0 ) is called a Clairaut equation (or Clairauts equation). Show
that the general solution of this equation is the family of straight lines y = cx + f (c). In addition
to these show that it has a special solution given by f 0 (p) = x where p = y 0 . This special solution
which does not (in general) represent one of the straight lines y = cx + f (c), is called a singular
solution. Hint: Differentiate the differential equation.
Q.13. Determine the general solutions as well as the singular solutions of the following Clairaut equations.
In each of the two examples, sketch the graphs of these solutions.
p
(i) y = xy 0 + 1/y 0 .
(ii) y = xy 0 y 0 / 1 + y 0 2

Q.14. For the parabola y = x2 find the equation of its tangent at (c, c 2 ) and find the ordinary differential
equation for this one parameter family of tangents. Identify this as a Clairaut equation. More
generally take your favourite curve and determine the ODE for the one parameter family of its
tangents and verify that it is a Clairuts equation. N.B: Exercise 13 shows that the converse is true.

Q.15. In the preceeding exercises, show that in each case, the envelope of the family of straight lines is also
a solution of the Clairaut equation.

Q.16. Show that the differential equation y 0 y 3 = 2x3/2 has three distinct solutions of the form A/ x
but that only one of these is real valued.

Tutorial Sheet No. 2

Q.1. State the conditions under which the following equations are exact.
(i) [f (x) + g(y)]dx + [h(x) + k(y)]dy = 0
(ii) (x3 + xy 2 )dx + (ax2 y + bxy 2 )dy = 0
(iii) (ax2 + 2bxy + cy 2 )dx + (bx2 + 2cxy + gy 2 )dy = 0
4

Q.2. Solve the following exact equations

(i) 3x(xy 2)dx + (x3 + 2y)dy = 0
(iii)

ex y(x

+ y)dx +

ex (x

(ii) (cos x cos y cot x)dx sin x sin ydy = 0.

+ 2y 1)dy = 0

Q.3. Determine (by inspection suitable) Integrating Factors (IFs) so that the following equations are
exact.
(i) ydx + xdy = 0
(ii) d(ex sin y) = 0
y
(iv) yex/y dx + (y xex/y )dy = 0
(iii) dx + ( )2 dy = 0
x
(v) (2x + ey )dx + xey dy = 0, (vi) (x2 + y 2 )dx + xydy = 0
Q.4. Verify that the equation M dx + N dy = 0 . . . (1) can be expressed in the form
1
1
x
(M x + N y)d(ln xy) + (M x N y)d ln( ) = 0.
2
2
y
Hence, show that (i) if M x + N y = 0, then
(ii) if M x N y = 0, then

1
is an IF of (1) and
Mx Ny

1
is an IF of (1).
Mx + Ny

Also show that (iii) if M and N are homogeneous of the same degree then

1
is an IF of (1).
Mx + Ny

Q.5. If (x, y) is an IF of M dx + N dy = 0 then prove that

My N x = N

ln || M
ln ||.
x
y

Z x
1
f (t)dt)
Use the relation to prove that if (My Nx ) = f (x) then there exists an IF (x) given by exp(
N
a
Z y
1
(My Nx ) = g(y), then there exists an IF (y) given by exp(
and if
g(t)dt). Further if
M
a
Rx
R
y
My Nx = f (x)N g(y)M then (x, y) = exp( a f (x0 )dx0 + a g(y 0 )dy 0 ) is an IF, where a is
any constant.
Determine an IF for the following differential equations:
(i) y(8x 9y)dx + 2x(x 3y)dy = 0.

(ii) 3(x2 + y 2 )dx + (x3 + 3xy 2 + 6xy)dy = 0

(iii) 4xy + 3y 2 x)dx + x(x + 2y)dy = 0

Q.6. Find the general solution of the following differential equations.
(i) (y xy 0 ) + a(y 2 + y 0 ) = 0
p
p
(iii) (x3 + y 2 x2 + y 2 )dx xy x2 + y 2 dy = 0

(v) y 0 x1 y = x1 y 2

(viii) y 0 + y tan x = cos2 x

(iv) (x + y)2 y 0 = 1

(vi) x2 y 0 + 2xy = sinh 3x

(ix) (3y 7x + 7)dx + (7y 3x + 3)dy = 0.

Q.7. Solve the following homogeneous equations.

p
p
(i) (x3 + y 2 x2 + y 2 )dx xy x2 + y 2 dy = 0,

(ii) (x3 + y 3 )dx 3xy 2 dy = 0

(iii) (x2 + 6y 2 )dx + 4xydy = 0,

(v) xy 0 = y + x cos2

(ii) [y + xf (x2 + y 2 )]dx + [yf (x2 + y 2 ) x]dy = 0

(iv) xy 0 = y(ln y ln x).

y
x

Q.8. Solve the following first order linear equations.

(i) xy 0 2y = x4

(iii) y 0 = 1 + 3y tan x

(ii) y 0 + 2y = e2x

(iv) y 0 = cosec x + y cot x.

(v) y 0 = cosec x y cot x.

(vi) y 0 my = c1 emx

Q.9. A differential equation of the form y 0 + f (x)y = g(x)y is called a Bernoulli equation. Note that if
= 0 or 1 it is linear and for other values it is nonlinear. Show that the transformation y 1 = u
converts it into a linear equation. Use this to solve the following equations.
dy
(i) ey y 0 ey = 2x x2
(iv)(xy + x3 y 3 )
= 1.
dx
dy
2
(ii) 2(y + 1)y 0 (y + 1)2 = x4
(v)
= xy + x3 y 3
x
dx
(iii) xy 0 = 1 y xy
(vi) xy 0 + y = 2x6 y 4
(vii) 6y 2 dx x(2x3 + y)dy = 0 (Bernoulli in x).

Q.10.

(i) Solve (x2 + 6y 2 )dx 4xydy = 0 as a Bernoulli equation.

(ii) Consider the initial value problem y 0 = y(1 y), y(0) = 0. Can this be solved by the meethod
of separation of variables? As a Bernoulli equation?
Put y = 1 u, u(0) = 1 and solve the resulting equation as a Bernoulli equation.

(iii) Solve 2ydx + x(x2 ln y 1)dy = 0. Hint: The equation is Bernoulli in x.

(iv) Solve cos y sin 2xdx + (cos 2 y cos2 x)dy = 0

(Hint: Put z = cos2 x; resulting ODE is Bernoulli in z.)

Q.11. Find the orthogonal trajectories of the following families of curves.

(i) x2 y 2 = c2

(v) y 2 = 4(x + h)

(ii) y = cex

(iii) ex cos y = c

(vi) y 2 = 4x2 (1 cx)

(viii) y = c(sec x + tan x).

(iv) x2 + y 2 = c2

(vii) y 2 = x3 /(a x)

(ix) xy = c(x + y)

(x) x2 + (y c)2 = 1 + c2
x2
y2
+
= 1,
a2 + b2 +
orthogonal trajectories. Explain the anomaly.

Q.12. Find the ODE for the family of curves

(0 < b < a) and find the ODE for the

Q.13. A differential equation of the form y 0 = P (x) + Q(x)y + R(x)y 2 is called Riccatis equation. In general, the equation cannot be solved by elementary methods. But if a particular solution y = y 1 (x)
is known, then the general solution is given by y(x) = y 1 (x) + u(x) where u satisfies the Bernoulli
equation
du
(Q + 2Ry1 )u = Ru2 .
dx
(i) Use the method to solve y 0 + x3 y x2 y 2 = 1, given y1 = x.

(ii) Use the method to solve y 0 = x3 (y x)2 + x1 y given y1 = x.

Q.14. Consider the differential equation y 0 + P (x)y = 0, where P (x) is continuous on an interval I. Show
that
(i) if y = f (x) is a solution and f (x0 ) = 0 for x0 I, then f (x) = 0 for all x I
6

(ii) if f (x) and g(x) are two solutions such that f (x 0 ) = g(x0 ) for some x0 I, then f (x) = g(x)
for all x I.
Q.15. Determine by Picards method, successive approximations to the solutions of the following initial
value problems. Compare your results with the exact solutions.

(i) y 0 = 2 y; y(1) = 0
(ii) y 0 xy = 1;

(iii)

y0

=x

y2;

y(0) = 1
y(0) = 1.

Q.16. Show that the function f (x, y) = | sin y| + x satisfies the Lipschitzs condition
|f (x, y2 ) f (x, y1 )| M |y2 y1 |
with M = 1, on the whole xy plane, but f y does not exist at y = 0.
Q.17. Examine whether the following functions satisfy the Lipschitz condition on the xy plane. Does
exist ? Compute the Lipschitz constant wherever possible.
(i) f = |x| + |y|

(ii) f = 2 y in < : |x| 1, 0 y 1 or in < : |x| 1,

1
2

f
y

<y<1

(iii) f = x2 |y| in < : |x| 1, |y| 1

(iv) f = x2 cos2 y + y sin2 x, |x| 1, |y| <

Tutorial Sheet No. 3

Q.1. Find the curve y(x) through the origin for which y 00 = y 0 and the tangent at the origin is y = x.
Q.2. Find the general solutions of the following differential equations.
(i) y 00 y 0 2y = 0 (ii) y 00 2y 0 + 5y = 0
Q.3. Find the differential equation of the form y 00 + ay 0 + by = 0, where a and b are constants for which
the following functions are solutions:
(i) e2x , 1

(ii) e(+i)x , e(i)x .

Q.4. Are the following statements true or false. If the statement is true, prove it, if it is false, give a
counter example showing it is false. Here Ly denotes y 00 + P (x)y 0 + Q(x)y.
(i) If y1 (x) and y2 (x) are linearly independent on an interval I, then they are linearly independent
on any interval containing I.
(ii) If y1 (x) and y2 (x) are linearly dependent on an interval I, then they are linearly dependent on
any subinterval of I.
(iii) If y1 (x) and y2 (x) are linearly independent solution of L(y) = 0 on an interval I, they are
linearly independent solution of L(y) = 0 on any interval I contained in I.
7

(iv) If y1 (x) and y2 (x) are linearly dependent solutions of L(y) = 0 on an interval I, they are linearly
dependent on any interval J contained in I.
Q.5. Are the following pairs of functions linearly independent on the given interval?
(i) sin 2x, cos(2x + 2 ); x > 0 (ii) x3 , x2 |x|; 1 < x < 1

(iii) x|x|, x2 ; 0 x 1 (iv) log x, log x2 ; x > 0 (v) x, x2 , sin x; x IR

Q.6. Solve the following:
(i) y 00 4y 0 + 3y = 0, y(0) = 1, y 0 (0) = 5;

(ii) y 00 2y 0 = 0, y(0) = 1, y( 12 ) = e 2.

Q.7. For what non-negative values of do there exist non trivial solutions of 00 + 2 = 0 satisfying
(i) (0) = 0 = (), (ii) 0 (0) = 0 = 0 ()
(iii) (0) = (), 0 (0) = 0 (), (iv)(0) = (), 0 (0) = 0 ().
Q.8. Solve the following initial value problems.
(i) (D 2 + 5D + 6)y = 0,
(iii) (D 2 + 2D + 2)y = 0,

y(0) = 2, y 0 (0) = 3

(ii) (D + 1)2 y = 0,

y(0) = 1, y 0 (0) = 2

y(0) = 1, y 0 (0) = 1

Q.9. Solve the following initial value problems.

(i) (x2 D 2 4xD + 4)y = 0, y(1) = 4, y 0 (1) = 1

(ii) (4x2 D 2 + 4xD 1)y = 0, y(4) = 2, y 0 (4) = 0.25

(iii) (x2 D 2 5xD + 8)y = 0, y(1) = 5, y 0 (1) = 18

Q.10. Using the Method of Undetermined Coefficients, determine a particular solution of the following
equations. Also find the general solutions of these equations.
(i) y 00 + 2y 0 + 3y = 27x
(iii) y 00 + 4y 0 + 4y = 18 cos hx
(v) y 00 + 4y 0 + 3y = sin x + 2 cos x
(vii) y 00 + y = x cos x + sin x
(ix) y 000 y 0 = 2x2 ex

(ii) y 00 + y 0 2y = 3ex
(iv) y 0000 + y = 6 sin x
(vi) y 00 2y 0 + 2y = 2ex cos x
(viii) 2y 0000 + 3y 00 + y = x2 + 3 sin x
(x) y 000 5y 00 + 8y 0 4y = 2ex cos x

Q.11. Solve the following initial value problems.

(i) y 00 + y 0 2y = 14 + 2x 2x2 , y(0), y 0 (0) = 0.

(ii) y 00 + y 0 2y = 6 sin 2x 18 cos 2x; y(0) = 2, y 0 (0) = 2.

(iii) y 00 4y 0 + 3y = 4e3x , y(0) = 1, y 0 (0) = 3.

Q.12. Find a solution y = y(x) of the initial value problem

00

y + y = f (x), y(0) = 0, y (0) = 1, where f (x) =

x
if 0 x
ex if x >

such that both y and y 0 are continuous functions.

Q.13. For each of the following equations, write down the form of the particular solution. Do not go further
and compute the Undetermined Coefficients.
(i) y +y = x3 sin x (ii) y +2y 0 +y = 2x2 ex +x3 e2x (iii) y 0 +4y = x3 e4x (iv) y (4) + y = xex/
00

00

sin(x/ 2).

00

00

Q.14. Solve the Cauchy-Euler equations: (i) x 2 y 2y = 0 (ii) x2 y +2xy 0 6y = 0. (iii) x2 y 00 +2xy 0 +y/4 =

1/ x
00

Q.15. Find the solution of x2 y xy 0 3y = 0 satisfying y(1) = 1 and y(x) 0 as x .

Q.16. Show that every solution of the constant coefficient equation y 00 + y 0 + y = 0 tends to zero as
x if and only if the real parts of the roots of the characteristic polynomial are negative.
Tutorial Sheet No. 4

Q.1. Using the Method of Variation of Parameters, determine a particular solution for each of the following.

(i) y 00 5y 0 + 6y = 2ex (ii) y 00 + y = tan x, 0 < x <

2

00
0
2 2x
00
(iii) y + 4y + 4y = x e , x > 0 (iv) y + 4y = 3 cosec 2x, 0 < x <
2
2
00
0
3
00
0
3
x
(v) x y 2xy + 2y = 5x cos x (vi) xy y = (3 + x)x e
Q.2. Let y1 (x) and y2 (x) be two solutions of the homogeneous equation y 00 + p(x)y 0 + q(x)y = 0, a < x < b,
and let W (x) be the Wronskian of these two solutions. Prove that W 0 (x) = p(x)W (x). If W (x0 ) = 0
for some x0 with a < x0 < b, then prove that W (x) = 0 for each x with a < x < b.
Q.3. Let y = y1 (x) be a solution of y 00 + p(x)y 0 + q(x)y = 0. Let I be an interval where y 1 (x) does not
vanish, and a I be any element. Prove that the general solution is given by
Rt
Z x
exp[ a p(u)du]
dt.
y = y1 (x)[c2 + c1 (x)] where (x) =
y12 (t)
a
Q.4. For each of the following ODEs, you are given one solution. Find a second solution.

(i) 4x2 y 00 + 4xy 0 + (4x2 1)y = 0; y1 (x) = sin x/ x

(ii) y 00 4xy 0 + 4(x2 2)y = 0; y1 = ex

(iii) x(x 1)y 00 + 3xy 0 + y = 0; y1 = x/(x 1)2 ;

(iv) xy 00 y 0 + 4x3 y = 0, y1 = cos x2

r
 3 x2 
1 x2
2
2 00
3 0
.
(v) x (1 x )y x y
y = 0, y1 =
4
x
(vi) x(1 + 3x2 )y 00 + 2y 0 6xy = 0, y1 = 1 + x2

(vii) (sin x x cos x)y 00 (x sin x)y 0 + (sin x)y = 0, y1 = x.

Q.5. Computing the Wronskian or otherwise, prove that the the functions e r1 x , er2 x , . . . , ern x , where
r1 , r2 , . . . , rn are distinct real numbers, are linearly independent.

Q.6. Let y1 (x), y2 (x) . . . , yn (x) be n linearly independent solutions of the nth order homogeneous linear differential equation y (n) + p1 (x)y (n1) + . . . + pn1 (x)y + pn (x)y = 0. Prove that y(x) = c1 (x)y1 (x) +
c2 (x)y2 (x)+. . .+cn (x)yn (x) is a solution of the nonhomogeneous
equation y (n) + p1 (x)y (n1) + . . . + pn1 (x)y = r
Z
Di (x)
dx, where Di (x) is the determinant of
where c1 (x), c2 (x), . . . , cn (x) are given by ci (x) =
W (x)
9

the
matrix
obtained from the matrix defining the Wronskian W (x) by replacing its i th column by
0
0

..
.
r(x)
Q.7. Three solutions of a certain second order non-homogeneous linear differential equation are
y1 (x) = 1 + ex

y2 (x) = 1 + xex ,

y3 (x) = (1 + x)ex 1.

Find the general solution of the equation.

Q.8. For the following nonhomogeneous equations, a solution y 1 of the corresponding homogeneous equation is given. Find a second solution y 2 of the corresponding homogeneous equation and the general
solution of the nonhomogeneous equation using the Method of Variation of Parameters.
(i) (1 + x2 )y 00 2xy 0 + 2y = x3 + x,

(ii) xy 00 y 0 + (1 x)y = x2 ,

y1 = x

(iii) (2x + 1)y 00 4(x + 1)y 0 + 4y = e2x ,

y1 = ex

y1 = e2x

(iv) (x3 x2 )y 00 (x3 + 2x2 2x)y 0 + (2x2 + 2x 2)y = (x3 2x2 + x)ex ,

y1 = x2

Q.9. Reduce the order of the following equations given that y 1 = x is a solution.
000

(i) x3 y 3x2 y 00 + (6 x2 )xy 0 (6 x2 )y = 0

000

(ii) y + (x2 + 1)y 00 2x2 y 0 + 2xy = 0

Q.10. Find the complementary function and particular integral for the following differential equations
(i) y (4) + 2y (2) + y = sin x (ii) y (4) y (3) 3y (2) + 5y 0 2y = xex + 3e2x
Q.11. Solve the following Cauchy-Euler equations
(i) x2 y 00 + 2xy 0 + y = x3 (ii) x4 y (4) + 8x3 y (3) + 16x2 y (2) + 8xy 0 + y = x3
(iii) x2 y 00 + 2xy 0 +

y
4

1
x

Q.12. Find a particular solution of the following inhomogeneous Cauchy-Euler equations.

(i) x2 y 00 6y = ln x (ii) x2 y 00 + 2xy 0 6y = 10x2 .
Q. 13. Find a second solution of
(i) (x2 x)y 00 + (x + 1)y 0 y = 0 given that (1 + x) is a solution.

(ii) (2x + 1)y 00 4(x + 1)y 0 + 4y = 0 given that e2x is a solution.

Q. 14. Find a homogeneous linear differential equation on (0, ) whose general solution is c 1 x2 ex + c2 x3 ex .
Does there exist a homogeneous differential equation with constant coefficients with general solution
c1 x2 ex + c 2 x3 ex ?

Tutorial Sheet No. 5

10

Q.1. Are the following real sequences bounded, eventually monotone and convergent?
(i) k, 2k 2 , 3k 3 . . . (|k| < 1)

(ii) a, 22 a2 , 32 a3 , 42 a4 , . . . (|a| < 1)

101 102
10n
1 13 135
,
,...
. . . . (iv) ,
,
,...
1! 2!
n!
2 24 246
1
1
1
1
1
(v) 1, 1 + , 1 + + , . . . , 1 + + . . . + , . . .
1!
1! 2!
1!
n!

 
1 1/n
converges to 1 as n .
Q.2. (i) Prove that ln 1 +
n
(iii)

(ii) Show that if {an } is monotone decreasing then so is the sequence

n1
n

a1 + a 2 + . . . + a n

o

Q.3. Examine the following series for convergence.

X
X
X
X
n!
1
1
(n!)2
(ii)
(iii)
(iv)
(i)
n
(2n)!
n
ln
n
ln(ln
n)
1
n=10
n=100
1

X
n1000

(v)

n=1

n!

(vi)

X
1 3 5 (2n 1)

n4

n=1

Q.4. Examine the series

n=0

(vii)

in
n2 + i
n=0

(viii)

X
n+i

2n n

n=1

1
1
an for convergence, where an = ( )n , if n is even and an = ( )n , if n is odd.
2
3

Q.5. Examine the following series for absolute and conditional convergence.

X
X
X
(1)n n n
1
(1)n
1
(ii)
(iii)
uk , where u2k = 2k and u2k+1 = 2k+1 , k = 0, 1, 2, . . .
(i)
2n
ln n
e +1
3
2
n=2

(iv)

n=1

k=0

(uk + vk ), where uk =

k=0

i
23k

and vk =

1
23k+1

, k = 0, 1, 2, . . .

Q.6. For what real values of x, are the following series convergent.

X
X
X
X
xn
1
xn
n
n
(i)
(x + x ) (iv)
(ii)
(iii)
n
n!
n
x + xn
(vi)

X
n=r

n(n 1) . . . (n r + 1)xnr

Q.7. (Modified Comparison Test.) Let

(vii)

n=1

un and

for n n0 . Show that

X
X
(i) if
vn is convergent, then so is
un .
Examine the series

X
1 3 5 (2n 1)
n=1

apply (i)).

2n (n + 1)!

Q.8. Discuss the convergence of (i)

n=1

(v)

n=0

xn+r
n(n + 1)(n + 2) (n + r)
X

(ii) if

vn be two series of positive terms. Let

X

un is divergent, then so is

for convergence. (Hint: Take vn =

(a1/n 1)

xn
1 + x2n

(ii)

n=1

11

(1 cos(1/n))

(iii)

vn+1
un+1

un
vn

vn .

1
and
(n + 1)(2n + 1)1/2

X
ln(1 + n1 )
, x>0
xln ln n
n=3

Q.9. Examine the convergence of the following improper integrals.

Z
Z 1
Z
Z 1
cos xdx
dx
dx
p1
q1

x
(1

x)
dx
(iv)
(i)
(ii)
(iii)
p+1
p+1
x
x
1 + x3
0
0
1
0
Z
x2m
(v)
dx (n, m 0)
1 + x2n
0

(iv)

xp1
dx
1+x

Q.10. Examine the convergence of the following improper integrals and determine the value of the convergent integrals.
Z
Z 1
Z
Z
dx
2
dx
x
n1
x

p
(i)
(iv)
e dx, (ii)
x
e dx, (iii)
1+2x2
x(1 x)
0
0
0
0
Q.11. Show that
by parts.

eax ebx
dx is convergent and find its value. Hint for the last part: Try integrating
x2

Q.12. (i) Show that both

(Fresnel Intergals)

sin x2 dx and

cos x2 dx are convergent integrals and find their values.

Z 2
dx
x dx
and
converge and find their values. Hint: Rewrite
4
1+x
1 + x4
0
0
them as beta integrals and use Eulers reflection formula.
Z a1
x dx
(iii) Express
as a beta integral and show that it is equal to B(a, 1 a) where 0 < a < 1.
1+x
0
(ii) Show that the integrals

Q.13. Test for convergence and evaluate the convergent integrals.

Z /2
Z 3
Z
dx
dx
dx

(i)
, (ii)
, (iii)
2
2
cos x
0
1 4x x 3
x + 2x + 5
Z /2

tan xdx.
(v)

(iv)

/2
0

dx.
tan x

Q.14. Test the following improper integrals for convergence:

Z
Z
Z
dx
dx
p
x sin x dx (ii)
(i)
(iii)
2
4
1 + 2x + 3x
x(1 x)(2 x)
0
0
3
Z 2
Z /2
2x 3
(iv)
dx (v)
ln sin x dx.
12
(x + 1)1/5
2
0
Z

a2 
Q.15. Prove that the improper integral
log 1 + 2 dx is convergent and that its value is a if a > 0.
x
0
Hint: Integrate by parts.
Q.16. Test the following series for convergence.

X
X
X
1
1
1
(ii)
(iii)
(i)
p
n ln n
n ln n
n(ln n)p
n=10

n=10

Q.17. Evaluate

10

exp(x2 )dx.

12

(iv)

n=100

1
.
n ln n(ln ln n)

Q.18. Discuss for convergence the following series:


X
X log(n!)
X
X
1
1
(1)n 
1
1
1
1 3 5 . . . (2n 1) p
(i)
1
+
(iv)
1
+
(iii)
(ii)
+
.
.
.
+
+
.
.
.
+
2n
2 4 6 . . . (2n)
n
2
n
n
2
n
n=1
n=1
n=1
X
X 1
1
(vi)
xn exp( log n)
(v)
1
2
nn+ n
Q.19. Show that the rearrangement

1 1 1 1 1 1
1
1
1+
+
+
+
+

3 2
5 7 4
9 11 6
converges and that its value is

3
2

log 2

Tutorial Sheet No. 6

Q.1. Examine each of the following sequence for uniform convergence in the given region.
x
, 0 < x < (iv) n2 xenx ,
(i) xn , 0 x 1 (ii) xn , 0 < x < 1 (iii)
1 + nx
(v) xn (1 x)n on [0, 1]

(0, )

Q.2. Show that if

n=1

(vi)

n2 x
on (0, )
1 + n 3 x2

(vii)

n2 x2
on (0, )
1 + n 3 x2

(viii)

0 x 1.
n2 x
on
1 + n 4 x2

fn (x) converges uniformly, then fn (x) 0 uniformly.

Q.3. Examine the uniform convergence of each of the following series

uk (x) in the given region.

(a) (i) uk =
(b)

X
1

x
,
((k 1)x + 1)(kx + 1)

(uk uk1 ) when (i) uk =

Q.4. Show that

n=0

0 < x < 1;

x
1 + k2 x

cosn x converges uniformly on

(ii) uk = xk1 xk , |x| < 1.

and (ii) uk =

kx
,
1 + k 2 x2

0 < x < .

 
 
. Is the convergence uniform on 0,
?
,
6 2
2

Q.5. Prove that

x2
is absolutely convergent but not uniformly convergent on [0, 1].
(1 + x2 )n1

Q.6. Show that

(1)n1
is uniformly convergent for all real x but not absolutely convergent for all
n + x2

real x.

13

Q.7. Prove that

X
(1)n (x2 + n)

converges uniformly in every bounded interval, but does not converge

n2
absolutely for any real value of x.
1

Q.8. Show that

X
1

x
converges uniformly and absolutely for all real x.
n(1 + nx2 )

n2 x
converges uniformly to zero on [0, 1]. What can you say about the
1 + n3 x
sequence of derivatives fn0 (x)?
Z 1
Z 1
n2 x
f
(x)dx
=
Q.10. Show that fn (x) =
lim fn (x)dx.
converges
to
zero
pointwise.
Examine
if
lim
n
n 0
1 + n 3 x2
0 n
Z /2
Z /2
2n
lim cos2n xdx although the sequence of functions cos 2n x
cos xdx =
Q.11. Show that lim
Q.9. Show that fn (x) =

n 0

fails to converge uniformly on [0, /2]. Hence determine the limit of the convergent sequence given
in tutorial sheet 5, Q1 (iv).
Hint: Break the integral into a sum of an integral over [0, ] and an integral over [, /2]. On [, /2]
we have uniform convergence and estimate the integral over [0, ].
Q.12. Prove that

X
1
1 X
,
converges uniformly if x 1 + > 1.
x
n
ln n(nx )
1

n=10

Q.13. Find the radius of convergence of the following power series:

X
X (2n)!
X
X zm
X
n
m
m(m 1) (m k + 1)z m (v)
(i)
z (ii)
(iii)
m!z (iv)
zn
m!
22n (n!)2
m=k

(vi)

X
1

zm
m(m + 1) (m + k + 1)

(vii)

X
1

X
nn n
(2n)! n
z (viii)
z
n!
nn
1

(ix)

Q.14. Suppose fn (x) converges uniformly to f (x) on (0, ) does it follow that
as n ?

Q.15. Determine the radius of convergence of

n!xn and

X
(3n)! n
z
2n (n!)3
1

fn (x)dx

f (x)dx

xn!

Tutorial Sheet No. 7

Q.1. Apply the power series method to determine the general solution of the following differential equations.
(i) (1 x2 )y 0 = y (ii) y 0 = xy y(0) = 1 (iii) (1 x2 )y 0 = 2xy
(iv) y 0 2xy = 1,

y(0) = 0. From (iv) deduce the Taylor series for e x

Q.2. Find the solution as a power series in powers of (x 1).

(i) y 00 + y = 0 (ii) y 00 y = 0

14

Rx
0

et dt.

Q.3. Find the power series solutions for the following differential equations.
(i) Legendres equation: (1 x2 )y 00 2xy 0 + p(p + 1)y = 0.

(ii) Tchebychevs equation: (1 x2 )y 00 xy 0 + p2 y = 0.

(iii) Airys equation: y 00 xy = 0.

(iv) Hermites equation : y 00 x2 y = 0.

Q.4. Show that the function (sin1 x)2 satisfies the IVP: (1 x2 )y 00 xy 0 = 2, y(0) = 0,
Hence, find the Taylor series for (sin 1 x)2 . What is its radius of convergence ?

y 0 (0) = 0.

Q.5. Attempt a power series solution (with center at the origin) for x 2 y 00 (1 + x)y = 0. Explain why the
procedure does not give any nontrivial solutions.
Q.6. Prove that if {f1 , . . . , fn , . . .} and {g1 , . . . , gn , . . .} are two sets of orthogonal vectors such that
o
o
n
n
Lin. Span f1 , . . . , fn = Lin. span g1 , . . . , gn , for all n,
then fn = cn gn for each n.
n
dn  2
x

1
for n = 1, 2, 3, . . . form an orthogonal family of polynomidxn
1 dn 2
als. Deduce from this the Rodrigues formula P n (x) = n
(x 1)n . Hint Use the preceeding
2 n! dxn
exercise.

Q.7. Prove that the polynomials

2 1/2

Q.8. Prove that (1 2xh + h )

Pn (x)hn

Q.9. Use the formula in Q.3, tutorial sheet 4 to find the second solution of the Legendre differential
equation for the cases p = 1 and p = 2.
Q.10. Show that if f (x) is a polynomial with double roots at a and b then f 00 (x) vanishes atleast twice in
(a, b). Generalize this and show (using Rodrigues formula) that P n (x) has n distinct roots in (1, 1)
and in particular, Pn (x) cannot have a double root.
Q.11. Establish the following recurrence relations for P n (x).
0
(ii) Pn+1
xPn0 (n + 1)Pn = 0

(i) (n + 1)Pn+1 (2n + 1)xPn + nPn1 = 0

0
(iii) xPn0 Pn1
nPn = 0

(v) (x2 1)Pn0 nxPn + nPn1 = 0

0
0
Pn1
(2n + 1)Pn = 0
(iv) Pn+1

Q.12. Prove the following relations:

(i) Pn (x) = (1)n Pn (x) (ii) Pn0 (x) = (1)n+1 Pn0 (x)
(iii) Pn (1) = 1; Pn (1) = (1)n (iv) P2n+1 (0) = 0; P2n (0) = (1)n
1
1
(v) Pn0 (1) = n(n + 1) (vi) Pn0 (1) = (1)n1 n(n + 1)
2
2
(2n
+
1)!
n
0 (0) = 0 (viii) P 0
.
(vii) P2n
2n+1 (0) = (1)
22n (n!)2

15

(2n)!
22n (n!)2

Q.13. Prove that

+1

Pm (x)Pn (x)dx = 0 (m 6= n) and

Pn (x)2 dx =

2
.
2n + 1

Q.14. Let z1 , z2 , . . . , zn be the zeros of Pn (x). Show that

n

1
1 1
1  X
1
+
=
+
2
2
2
Pn (x) (1 x )
2 1x 1+x
(x zj ) (1 zj2 )(P 0 (zj ))2
j=1

Hence determine the second solution of the Legendre equation of order n using the formula
Z
 Z

1
exp

P
(x)dx
dx
y2 = y 1
y12
Q.15. Prove the following relation if (n m) is even (m n)
Z +1
Z 1
0
0
(i)
Pm Pn dx = m(m + 1) (ii)
Pm Pn dx = 0 if (n m) is even and n 6= m.
1

(iii)

xm Pn0 (x)dx = 0 if m n. What is the value of the integral if instead n m is odd?

Tutorial Sheet No. 8

Q.1. Locate and classify the singular points for the following differential equations:
(i) Bessels equation: x2 y 00 + xy 0 + (x2 p2 )y = 0.

(ii) Laguerres equation: xy 00 + (1 x)y 0 + y = 0.

(iii) Jacobis equation: x(1 x)y 00 + ( ( + 1)x)y 0 + n(n + )y = 0.

(iv) The hypergeometric equation: x(1 x)y 00 + [c (a + b + 1)x)]y 0 aby = 0.

(v) xy 00 + (cot x)y 0 + xy = 0

Q.2. Attempt a Frobenius series solution y(x) = x

n=0

for the x2 y 00 + (3x 1)y 0 + y = 0, and compute

the successive coefficients and the radius of convergence of the series solution. Why does the method
fail?
Q.3. Find two linearly independent solutions of the following differential equations:
(i) x(x 1)y 00 + (4x 2)y 0 + 2y = 0.

(iii) x2 y 00 + x3 y 0 + (x2 2)y = 0.

(ii) 2x(x + 2)y 00 + y 0 xy = 0.

(iv) xy 00 + 2y 0 + xy = 0.

Q.4. Show that the hypergeometric equation has a regular singular point at infinity 1 , but that the point
of infinity is an irregular singular point for the Airys equation.
1
The differential equation y 00 + p(x)y 0 + q(x)y = 0 has a regular singular point at infinity, if after substitution of x = 1/t
in the ODE, the resulting ODE has a regular singular point at the origin.

16

Q.5. Using the indicated substitutions, reduce the following differential equations to Bessels equation and
find the general solution in term of the Bessel functions.
(i) x2 y 00 + xy 0 + (2 x2 2 )y = 0, (x = z). (ii) xy 00 5y 0 + xy = 0, (y = x3 u).

2
(iii) y 00 + k 2 xy = 0, (y = u x, kx3/2 = z).
3
2 00
0
2 2
(iv) x y + (1 2)xy + (x + 1 2 )y = 0, (y = x , x = z).
Q.6. (a) Prove that [xn Jn ]0 = xn Jn1 and [xn Jn ]0 = xn Jn+1
2n
(b) Use (a) to prove that (i) Jn1 + Jn+1 =
Jn
(ii) Jn1 Jn+1 = 2Jn0
x
r
r 
r

2
2
2 sin x
sin x (ii) J1/2 =
cos x
(iii) J3/2 =
cos x
Q.7. Show that (i) J1/2 =
x
x
x
x
Q.8. When n is an integer show that
(i) Jn (x) is an even function if n is even

(ii) J n (x) is an odd function if n is odd.

Q.9. Show that between any two consecutive positive zeros of J n (x) there is precisely one zero of Jn+1 (x)
and one zero of Jn1 (x).

 X

tx
x
Q.10. Prove that exp
Jn (x)tn (This formula is due to Schlomilch). Use Schlomilchs for=

2
2t

a
X
mula to show that J02 + 2
Jn2 = 1. Deduce that |J0 | 1; |Jn | 12 .
1

Q.11. Prove that (i) cos(x sin ) = J0 (x) + 2

cos 2nJ2n (x) (ii) sin(x sin ) = 2

sin(2n + 1)J2n+1 (x).

n
n+1 2
d
1 d 2
2
2
[Jn + Jn+1
] = Jn2
Jn+1 ,
[xJn Jn+1 ] = x(Jn2 Jn+1
), and deduce that
2 dx
x
x
dx

i
X
X
x
dh X
2
2
(2n + 1)Jn Jn+1
(2n + 1)Jn Jn+1 = . Hint for (ii): Look at
x
Jn = 1 (ii)
(i) J0 + 2
2
dx n=0
0
1

Q.12. Show that

Q.13. Prove the following.

000

(i) J3 + 3J00 + 4J0 = 0

(ii) J2 J0 = aJc00 find a and c.
Z
Z
(iii)
J+1 dx = J1 dx 2J .
Q.14. Using the identity Jp+1 = Jp0 + pJp /x, to prove that
r
2 n+ 1  1 d n  sin x 
n
(i) Jn+ 1 (x) = (1)
(Use induction)
x 2
2

x dx
x
2p
1
Jp1 (x)
=
2p+2
(ii)
What happens when p = 1/2? (proceed formally)
1
Jp (x)
x
x 2p+4 ... .
x

17

Q.15. Use the Schlomilch formula and the identity e x(t t )/2 ey(t t )/2 = e(x+y)(t t )/2 to prove (for n =

X
Jk (x)Jnk (y)
0. 1, 2, . . .) Jn (x + y) =
k=

Q.16. Show that

0.

i
h
2
2
x c1 J1/3 ( x3/2 ) + c2 J1/3 ( x3/2 ) is the general solution of the Airys equation y 00 +xy =
3
3

Q.17. Find two linearly independent series solutions of Bessels equation with p = 0. (note that one of these
will not be a power series). Perform a similar exercise for Laguerres equation.

Additional Problems involving the Bessel and Legendre functions:

1. Laplaces integral for Pn (x). The following integral representations were given by Pierre Simon
Laplace in his Mecanique Celeste. Prove the following:
Z
p
1
(x + x2 1 cos )n d
(i) Pn (x) =
0
Z
1
d

(ii) Pn (x) =
0 (x + x2 1 cos )n+1
2. A Frobenius series solution (x 1)

n=0

an (x 1)n for Legendres differential equation (1 x 2 )y 00 2xy 0 + p(p +

is sought. Determine the indicial equation and the Frobemius exponents. Show that there is a power
series solution that converges in the disc |x 1| < 2.
Z x
cos(x t)J1 (t)dt.
3. Show that xJ0 (x) satisfies y 00 + y = J1 (x) and hence deduce that xJ0 (x) =
0

4. Prove that for a non-negative integer n, J k (x) =

cos(x sin k)d

Hint: Put t = ei in Schlomilchs formula and integrate (after multiplying through by e ik ).

5. Prove the following estimate used in the proof of Schlomilchs formula: |J p (x)|

|x|p |x|
e .
p!

6. Prove the following for integral values of p:

Z
1 1 itx
e (1 t2 )1/2 dt
(i) J0 (x) =
1
Z 1
xp
(ii) Jp (x) = p
eitx (1 t2 )p1/2 dt
2 (p + 12 ) 1
7. Show that xy 00 + y 0 + xy = 0 has a solution of the form (ln x)Q 1 (x) + Q2 (x), where Q1 (x) and Q2 (x)
are power series. Hint: One solution is a power series with non-zero constant term. Obtain other
solution using the formula in Q.3, tutorial sheet 4.
8. For a function f (x1 , x2 ) which decays sufficiently rapidly the Fourier transform is defined as
Z

f(1 , 2 ) =
f (x1 , x2 ) exp(i(x1 1 + x2 2 ))dx1 dx2 .
R2

18

p
(i) Show that if f (x1 , x2 ) is radial, that is it depends only on
r
=
x21 + x22 , then so is its Fourier
p
1 , 2 ) depends only on = 2 + 2 ).
transform (that is, show that f(
1
2

(ii) Let f (x1 , x2 ) = (r) and f(1 , 2 ) = (). Then

() = 2

r(r)J0 (r)dr.

Solution to Q1(i): The integral is obviously a polynomial since upon expanding the integrand we find that
the integrals of odd powers of the cosine vanish. Call the integrals Q n (x). Then clearly
Q0 (x) = 1 = P0 (x),

Q1 (x) = x = P1 (x).

Our job will be over if we establish the relation

(n + 1)Qn+1 (2n + 1)xQn (x) + nQn1 (x) = 0.

For convenience we shall denote x + x2 1 cos by A. Now writing An+1 = xAn + An x2 1 cos we
get

Z
x2 1 n1
A
cos d
Qn (x) = xQn1 (x) +

0
Upon integrating by parts once we get
(n 1)(x2 1)
Qn (x) = xQn1 (x) +

An2 sin2 d

()

Write (x2 1) sin2 as follows:

p
x2 1 cos )2
p
p
= (x2 1) [(x + x2 1 cos )2 x2 2x x2 1 cos ]
p
p
= (x2 1) [A2 x(x + x2 1 cos ) x x2 1 cos ]

(x2 1) sin2 = (x2 1) (

= (x2 1) [A2 xA x(A x)] = 1 A2 + 2xA.

Substituting in (*) we get

Qn (x) = xQn1 (x) + (n 1)(Qn2 (x) Qn (x) + 2xQn1 (x)),
which upon rearrangement gives
nQn (x) (2n 1)xQn1 (x) + (n 1)Qn2 (x) = 0
Replacing n by n + 1 we get the relation sought. The second formula of Laplace can be derived similarly.

19

Tutorial Sheet No. 9

Q.1. Solve the following boundary value problems.

(i) y 00 y = 0, y(0) = 0, y(1) = 1

(ii) y 00 6y 0 + 25y = 0, y 0 (0) = 1, y(/4) = 0

(iii) x2 y 00 + 7xy 0 + 3y = 0, y(1) = 1, y(2) = 2

(iv) y 00 + y 0 + y = x, y(0) + 2y 0 (0) = 1, y(1) y 0 (1) = 8

(v) y 00 + 2 y = 0, y(1) = y(1), y 0 (1) = y 0 (1).

Q.2. Find the eigenvalues and eigenfunctions of the following boundary value problems.
(i)y 00 + y = 0,

y 0 (1) = 0.

y(0) = 0,

00

(ii)y 00 + y = 0,

(iii)y + y = 0, y(0) = y (0), y(1) = 0.,

(v)(e2x y 0 )0 + e2x ( + 1)y = 0,

y(0) = 0,

y(0) = 0,

00

(iv)y + y = 0,

y(`) = 0.

y(0) = y(2), y 0 (0) = y 0 (2).

y() = 0.

Q.3. For which values of , does the boundary value problem

y 00 2y 0 + (1 + )y = 0,

y(0) = 0, y(1) = 0

have a non-trivial solution ?

Q.4. Show that the eigenvalues of the boundary value problem
y 00 + y = 0, y(0) = 0, y(1) + y 0 (1) = 0

are obtained as solutions of tan k = k, where k = . Conclude from a plot that this equation has
infinitely many solutions. Show that the eigenfunctions are y m = sin(km x).
Q.5. Determine the normalised eigenfunctions of the Sturm-Liouville problem y 00 + y = 0, y(0) = 0 = y(1).
Q.6. Expand the function f (x) = x, x [0, 1] in terms of the normalised eigenfunctions n (x) of the
boundary value problem y 00 + y = 0, y(0) = 0, y(1) + y 0 (1) = 0.
Q.7. Find the eigenfunctions and the eigenvalues of the following Sturm-Liouville problems.
(i) y 00 + 2y 0 + ( + 1)y = 0;

y(0) = y() = 0
(ii) x2 y 00 + xy 0 + y = 0; y(1) = y(`) = 0.
  2



d
k x n2
d
kx
kx
)
satisfies

)
+
Q.8. Verify that Jn ( kx
x
J
(
Jn ( ) = 0.
n
a
dx dx
a
a2
x
a
Multiply by Jn ( `x
a ) and integrate by parts from 0 to a to get
kJn0 (k)Jn (l)

kl
2
a

a
0

xJn0

`x
a

Jn0

kx
a

dx +

(
0

k2
n2
kx
lx
x

)Jn ( )Jn ( )dx = 0,

2
a
x
a
a

where prime (0) denotes differentiation with respect to the argument of J n . Interchange k and ` to
obtain the relation
Z a
`x
`Jn (k)Jn0 (`) kJn (`)Jn0 (k)
kx
.
xJn ( )Jn ( )dx = a2
a
a
k 2 `2
0

20

Prove that if k and ` are the roots of the Bessels equation J n () = 0 then
Z a
kx
`x
xJn ( )Jn ( )dx = 0
(k 6= `)
a
a
0
1 2 0
=
a [Jn (k)]2
(k = `)
2
1 2 2
=
a Jn+1 (k)
(k = `).
2
d
Q.9. The function Pn (x) satisfies the equation
[(1 x2 )Pn0 ] + n(n + 1)Pn = 0. Proceed as indicated in
dx
Z +1
Q.8 above to prove that
Pm Pn dx = 0
(m 6= n).
1

Q.10. If xn =

n
X

ar Pr (x) prove that an =

r=0

Q.11. Prove that

+1

2n (n!)2
.
(2n)!

(1 x2 )[Pn0 (x)]2 dx =

2n(n + 1)
.
2n + 1

Q.12. Represent the following functions in terms of Legendre Polynomials: (i) 5x 3 +x (ii) 10x3 3x2 5x1
Q.13. Show that f (x) = xn , (0 < x < 1; n = 0, 1, 2, . . .) can be represented by the Fourier-Bessel series
X 2Jn (ki x)
where ki s are the roots of Jn (k) = 0.
xn =
ki Jn+1 (ki )
Q.14. Represent the following functions in a Fourier-Bessel series containing the functions J 0 (ki x/a) where
ki are the roots of J0 (k) = 0. (i) f (x) = a2 x2 (0 < x < a) (ii) f (x) = 1 (0 < x < a/2); f (x) =
0, (a/2 < x < a).

Tutorial Sheet No. 10

Q.1. Show that

X
1
sin nx sin2 n = constant
n
1
= 0

(0 < x < 2)

(2 < x < )

1 (0 < x < 1/3)

X
1 4
1
2
1 (2/3 < x < 1)
Q.2. Prove that +
.
sin m cos 2mx =

3 1 m
3
1 (1/3 < x < 2/3)

Q.3. Show that

X
1
(2n + 1)x
( 2x)( 2 + 2x 2x2 ) =
cos
96
(2n + 1)4
0

Q.4. Prove that

X
(1)n1 cos nx
1

n2

2 x2

12
4

( x ).
21

(0 x ).

Q.5. Show that

X
sin(2n + 1)x
0

Q.6. Prove that

X
cos(2n + 1)x
0

Q.7. Show that

2n

+ 1)3

(2n +

1)2

1
= x( x)
8

(0 x ).

( 2x)
8

(0 < x < ).

1 1
cos 2x cos 3x cos 4x
1
+ cos x

. . . = ( x) sin x(0 x ).
2 4
1.3
2.4
3.5
2

Q.8. (Fourier Theorem.) Let f (x) be a periodic function

of period 2 on the real axis which is piecewise
Z

continuously differentiable. Suppose further that

relations

|f (x)|dx < . Let an and bn be defined by the

Z
Z
1
1
f (t) cos ntdt and bn =
f (t) sin nt dt, n = 0, 1, 2, . . .


P
The series 12 a0 +
1 (an cos nx + bn sin nx) converges to f (x) if f (t) is continuous at t = x and
1
converges to 2 [f (x + 0) + f (x 0)] if f (t) has a finite discontinuity at t = x. From the Fourier
expansions given in Q.1 through Q.7 and the Fourier theorem stated above deduce the following
results.
1
1
2
1
1
1 1 1 1 1
1 1 1 1 1

+ ... =

+ ...
(ii) 1 + + +
(i) 1 + + +
2 4 5 7 8 10 11
2 4 5 7 8 10 11
3 3

=
3 3
1
1
1
2
1
1
1
2
(iv) 1 2 + 2 2 + . . . =
(v) 1 + 2 + 2 + 2 + . . . =
2
3
4
12
2
3
4
6
3
1
1
1

1
1
1
2
1
(vii) 1 + 2 + 2 + 2 . . . =
(vi) 1 3 + 3 3 + 3 + . . . =
3
5
7
9
32
3
5
7
8
1
1
1
1
1
(viii)

+ ... =
1.3 3.5 5.7 7.9
4 2
an =

Q.9. Using Parsevals identity, prove that 1 + 314 + 514 +

x,
/2 < x < /2

(Hint: Use f (x) =

x,
/2 < x < 3/2.

1
74

+ ... =

4
96 .

Q.10. Find the Fourier series of the function f (x) which is assumed to have the period 2, where
(i) f (x) = x, 0 < x < 2.

x < 0
x,
(ii) f (x) =

x,
0x<

(iii) f (x) = x + |x|,

< x < .

Q.11. Find the Fourier series of the periodic function f (x), of period p = 2, when

1 < x < 0
0,
f (x) =

x,
0<x<1
22

Q.12. State whether the given function is even or odd. Find its Fourier series

/2 < x < /2
k,
(i)f (x) =
(ii) f (x) = 3x( 2 x2 ), < x < .

0,
/2 < x < 3/2
Q.13. Using Fourier Integrals, show that

Z
if x < 0
0
cos xw + w sin xw
/2
if
x=0
dw
=
(i)

1 + w2
x
0
e
if x > 0
(iii)

w3 sin xw

dw = ex cos x(x > 0)

4
w +4
2

Q.14. Let f (x) be defined over R and

A(w) =
The Fourier integral

(ii)

(iv)

cos xw

dw = ex (x > 0)
2
1+w
2
0

/2 if 0 x < 1
sin w cos xw
/4 if
x=1
dw =

w
0 if
x > 1.

|f (x)|dx < . Further, let

f (v) cos wvdv; B(w) =

f (v) sin wvdv

(A(w) cos wx + B(w) sin wx)dw converges to f (x) if f (t) is continuous at

t = x and converges to 12 [f (x + 0) + f (x 0)] if f (t) has a finite discontinuity at t = x. From the

Fourier integral theorem stated above deduce the following results.
Z
Z
Z 3
Z
w sin w
cos w
w sin w

sin w
2
dw = /2
(ii)
dw
=
(iii)
dw
=
(iv)
dw = 0
(i)
w
1 + w2
2e
1 + w2
2e
w4 + 4
0
0
0
0
Z
2
Q.15. If A(w) =
f (v) cos wv dv, then show that
0
Z
Z
Z 2
1 w
dA
d A
2
(i) f (ax) =
cos wx dw.
sin wx dw. (iii) x f (x) =
A( ) cos wx dw. (ii) xf (x) =
a 0
a
dw
dw2
0
0
x
1
, and the Fourier sine integral of f (x) =
.
2
1+x
1 + x2
Both functions are defined on the interval [0, ).

Q.16. Find the Fourier cosine integral of f (x) =

Tutorial Sheet No. 11

Q.1. Find the Laplace Transform of the following functions.

(i) t cos wt (ii) t sin wt (iii) et sin2 t (iv) t2 eat (v) (1 + tet )3 (vi) (5e2t 3)2
(vii) te2t sin wt (viii) tn eat (ix) t2 eat sin bt (xi) cosh at cos at

Q.2. Find the inverse Laplace transforms of the following functions.

(i)

s2 w 2
2as
1
s3
s2
1
1
(ii)
(iii)
(iv)
(v) 2
(vi) 4
(vii) 4 2
(s2 + w2 )2
(s2 a2 )2
(s2 + w2 )2
(s4 + 4a4 )
s (s + 4)2
s 2s3
s (s + 2 )

(viii)

s3 + 3s2 s 3
s3 7s2 + 14s 9
s2 + a 2
(ix)
(x)
(s2 a2 )2
(s2 + 2s + 5)2
(s 1)2 (s 2)2
23

Q.3. Solve the following intial value problems using Laplace transforms and convolutions.
(i) y 00 + y = sin 3t; y(0) = y 0 (0) = 0 (ii) y 00 + 3y 0 + 2y = et ; y(0) = y 0 (0) = 0
(iii) y 00 + 2y 0 8y = 0; y(0) = 1; y 0 (0) = 8 (iv) y 00 + 2y 0 + y = 2 cos t; y(0) = 3, y 0 (0) = 0
(v) y 00 2y 0 + 5y = 8 sin t 4 cos t;
(vi) y 00 2y 0 3y = 10 sin h2t;

y(0) = 1; y 0 (0) = 3

y(0) = 0; y 0 (0) = 4

Q.4. Solve the following systems of differential equations using Laplace transforms.
(i) x0 = x + y, y 0 = 4x + y

(ii) x0 = 3x + 2y, y 0 = 5x + y

(iii) x00 x+y 0 = y = 1, y 00 +y+x0 x = 0

(iv) x0 = 5x+8y+1, y 0 = 6x9y+t, x(0) = 4, y(0) = 3

(v) y10 + y2 = 2 cos t; y1 + y20 = 0; y1 (0) = 0; y2 (0) = 1

(vi) y100 + y2 = 5 cos 2t; y200 + y1 = 5 cos 2t; y1 (0) = 1, y10 (0) = 1, y2 (0) = 1, y20 (0) = 1
(vii) 2y10 y20 y30 = 0; y10 + y20 = 4t + 2; y20 + y3 = t2 + 2, y1 (0) = y2 (0) = y3 (0) = 0
(viii) y100 = y1 + 3y2 ; y200 = 4y1 4et ; y1 (0) = 2; y10 (0) = 3, y2 (0) = 1, y20 (0) = 2
1
s

with no constant term the Laplace transform can be obtained

X
X
Ak
tk
1
term-by-term, i.e., assuming that L [
]
=
, where A0 , A1 . . . Ak . . . are real numbers,
A
k
sk+1
k!
0
0
prove that
1
1
) = et
(ii) L1 ( 2
) = sin t
(i) L1 (
s1
s +1

1
1
(iii) L1 ( eb/s ) = J0 (2 bt) (b > 0)
(iv) L1 (
) = J0 (at) (a > 0)
2
s
s + a2

sin t
1
1
eb/s
(v) L1 ( ) = cos(2 bt) (b > 0) (vi) L1 (tan1 ) =
s
s
t
t

Q.5. Assuming that for a Power series in

Q.6. Find the Laplace transform of the following periodic functions.

(i) f (t), f (t + p) = f (t) for all t > 0 and f (t) piecewise continuous
(ii) f (t) = | sin wt|

(iii) f (t) = 1(0 < t < ); f (t) = 1( < t < 2); f (t + 2) = f (t)
(iv) f (t) = t(0 t 1), f (t) = 2 t(1 t 2); f (t + 2) = f (t)

(v) f (t) = sin t(0 t ), f (t) = 0( t 2); f (t + 2) = f (t)

Q.7. Find the Laplace Transform of f (t) where f (t) = n, n 1 t n, n = 1, 2, 3, . . .

Q.8. Find f (t) given L[f (t)] = (es e2s e3s + e4s )/s2
Q.9. Find the Laplace Transform of (i) f (t) = u (t) sin t (ii) f (t) = u1 (t)e2t where u (u1 ) is the Heaviside
step function.


s2 + 4s + 5
1
ln 2
Q.10. Find (i) L
s + 2s + 5
Z t
1
Q.11. If L[f (t)] = F (s), L[g(t)] = G(s) prove that L [F (s)G(s)] =
f (u)g(t u)du. Also show that
0

L1 [

1
F (s)
] = eat
(s + a)2 + a2
a

f (u)eau sin a(t u)du.

24

Q.12. Compute
the Laplace transform of a solution of ty 00 + y 0 + ty = 0, t > 0, satisfying y(0) = k,
Y (1) = 1/ 2, where k is a real constant and Y denotes the Laplace transform of y.
Q.13. Compute the convolution of ta1 u(t) and tb1 u(t) and use the convolution theorem to prove
(a)(b) = (a + b)B(a, b)
where B(a, b) denotesZ the Beta function and (a) the Gamma function. Use this to find the value of

exp(x2 )dx.
(1/2) and hence of

Q.14. Suppose f (x) is a function of exponential type and Lf = 1/

s2 + 1. Determine f f.

Q.15. Evaluate the following integrals by computing their Laplace transforms.

Z
Z
Z
cos tx
sin(tx)
dx
(ii) f (t) =
dx (iii) f (t) =
(i) f (t) =
sin(txa )dx,
2 + a2
x
x
0
0
0
Z
Z
Z  2

4
2
1
sin tx
x b sin tx
(iv)
(1 cos tx)dx (v)
dx (vi)
dx
2
3
2 + b2
x
x
x
x
0
0
0
Q.16. Solve the following integral/integro-differential equations
Z
Z t
(1 + x)y(t x)dx
(ii) A =
(i) y(t) = 1 sinh t +
0

dy
=1
(iii)
dt

t
0

a>1

y(x)dx

, where A is a constant.
tx

y(t )d, y(0) = 1.

Q.17. Find a real general solution of the following nonhomogeneous linear systems.
(i) y10 = y2 + e3t ,

y20 = y1 3e3t .

(ii) y10 = 3y1 + y2 3 sin 3t,

(iii) y10 = y2 + 6e2t ,

y20 = 7y1 3y2 + 9 cos 3t 16 sin 3t.

y20 = y1 3e2t ,

y1 (0) = 11,
0
0
(iv) y1 = 5y2 + 23, y2 = 5y1 + 15t, y1 (0) = 1,
(v) y10 = y2 5 sin t, y20 = 4y1 + 17 cos t, y1 (0)
(vi) y10 = 5y1 + 4y2 5t2 + 6t + 25, y20 = y1 + 2y2

y2 (0) = 0.

y2 (0) = 2.

= 5,

y2 (0) = 2.

t + 2t + 4,

y1 (0) = 0,

y2 (0) = 0.

Q.18 Prove that the Laplace transform of (1 e t ) is B(s, + 1) where B(a, b) is the beta function.
Q.19 Show that if f (t) = 1/(1 + t2 ) thenZ its Laplace transform F (s) satisfies the differential equation

sin d
F 00 + F = 1/s. Deduce that F (s) =
.
( + s)
0
Q.20 Show that the Laplace transform of log t is s 1 log s Cs1 . Identify the constant C in terms of
the gamma function.
Z
n 
b o dt
where a and b are positive. Use this result to
Q.21 Evaluate the integral
exp at +
t
t
0

b 
1
., b > 0.
compute the Laplace transform of exp
t
t
Tutorial Sheet No. 12

25

Q.1. Determine the orbit of each of the following systems:

(i)x = y, y = x

(iii) x = 2xy, y = x 2 y 2

(ii)x = y(1 + x + y), y = x(1 + x + y)

(iv) x = xye3x , y = 2xy 2

(v) x = ax bxy, y = cx dxy, (a, b, c, d > 0).

Q.2. Find the equilibrium points of the following differential equations

(i) x 1 = 1 x2 ,

x 2 = x31 + x2 .

(ii) x 1 = x21 + x2 1,

x 2 = 2x1 x2 .

(iii) x 1 = (x1 1)(x2 1), x 2 = (x1 + 1)(x2 + 1).

(iv) y + y (y 3 + y 2 2y) = 0.

(v) x 1 = x22 5x1 + 6, x2 = x1 x2 .

Q.3. Describe the phase portrait of each of the following system and determine the nature and stability
properties of the critical points.
(i) x1 = x1 ,

(iii) x1 = x2 ,

x2 = x2 .
x2 = x1 .

(v) x1 = x1 2x2 ,

(ii) x1 = x1 ,

x2 = x1 + 2x2 .

(iv) x1 = ax1 x2 ,

x2 = x1 + ax2 ,

x2 = 4x1 5x2 . (vi) x1 = 4x1 x2 ,

x2 = x1 2x2 .

a 6= 0.

= AX,
where A =
Q.4. Determine whether each solution of the system of differential equations X

0
2
1
2 3
0
1
0
0

(i)
2
(ii) 0 6 2 (iii) 1 3 1 .
2 1

1
1 1
6
0 3
3 2 1
is stable, asymptotically stable or unstable.

Q.5. Discuss qualitative behaviour of the equilibrium points of Q.2 (i)-(v) and (viii)
( < 2, = 2, > 2) .
Q.6. (Competing species) Each of the following problems (i) - (vi) can be interpreted as describing the
interaction of two species with population x and y. In each of these problems carry out the following
steps:
(a) Find the equilibrium or the critical points.
(b) For each critical point find the corresponding linear system. Find the eigenvalues and eigenvectors of the linear system. Classify each critical point and determine whether it is asymptotically
stable, stable or unstable.
(c) Determine the limiting behaviour of x and y as t and interpret the result in terms of the
populations of two species.


(i) x = x(1 x y), y = y 43 y 21 x


(ii) x = x(1 x y), y = y 21 14 y 34 x .
(iii) x = x( 23 12 x y), y = y(2 y 1.125x)
(iv) x = x(1 x + 21 y), y = y(2.5 1.5y + 0.25x)
(v) x = x(1.5 0.5y), y = y(0.5 + x).
Q.7. For the following problems, carry out the steps in 6(a) and (b) and also discuss the limiting behaviour
of the solution.
(i) The Predator-Prey model: x = ax bxy, y = cy + dxy;
(a, b, c, d, positive real constants).
26

(ii) The Lanchestrian model for conventional guerilla war:

x = ay,
y = bx cxy.

27

Handout (Improper Integrals)

In Riemann integration the range of integration is finite and the integrand is bounded in that range.
It is possible, however, to so extend the theory that the symbol
Z

f (x) dx

(1)

may sometimes have a meaning even when either a or b or both are infinite or f (x) is not bounded. It is
convenient to have a local definition of boundedness of a function. A function f : (a, b) 7 IR is said to be
bounded at a point c in (a, b) if it is bounded in some interval (c , c + ) around c which is contained in
(a, b), otherwise, it is said to be unbounded at c. The function f is said to be unbounded at the left end
point a (respectively at the right end point b), if it is unbounded in every interval (a, a + ) (respectively,
(b , b)) contained in (a, b). If either a or b or both are infinite and f is bounded at each point contained
in (a, b) and at a or b, whenever they are finite, then the symbol (1) is called
Z integral of
Z an improper
dx
ex dx and
,
the first kind. Some examples of improper integrals of the first kind are
2
1
+
x
0

cos x dx.

If a and b are both real, but f is unbounded at some point in [a, b], then (1) is called an imZ 1
sin x
proper integral of the second kind. Examples of improper integrals of the second kind are
,
5/2
0 x
Z 1
Z 3
dx
dx
and
.
1/2
1 x
2 (x 2)(x 3)
If the integral of a function f is to be evaluated over an unbounded interval (a, b) and also f is
unbounded at some point of [a, b], then (1) is called an improper integral of the third kind. Some
examples of improper integrals of the third kind are
Z
Z
dx
dx
(i)
.
(ii)
3
1/3
(1 x)
x +x
0
0
Convergence of improper integrals of the first kind.
Definition. Let f be Riemann integrable in [a, R] for every R > a. 2 Then the improper integral
Z
Z R
f (x)dx of first kind is said to be convergent if lim
f (x)dx exists and is finite. The improper
R a
a
Z
integral
f (x)dx is said to be divergent, if it is not convergent.
a
Z
dx
, a > 0, for all real values of p.
Example 1. Discuss the convergence of
xp
a
Z R
R1p a1p
dx
Case I (p < 1)
lim
=
lim
= (divergent)
R a xp
R
1p
Z R
dx
Case II (p = 1)
lim
= lim (log R log a) = (divergent)
R a
R
x
Z R
a(1p)
dx
R(1p) a(1p)
Case III (p > 1)
lim
=
lim
=

finite (convergent).
R a xp
R
1p
1p

In place of this hypothesis, we may simply assume that f is continuous in the interval a x < .

28

Example 2. Show that

cos x dx is divergent.

Since lim

R 0

cos x dx = lim sin R does not exist, the integral is divergent.

R

Theorem 1 (The comparison test). Assume that

(i) f (x) and g(x) are Riemann integrable in [a, R] for every R > a,
(ii) 0Z f (x) g(x), a x < and
(iii)
Z

Then

g(x)dx is convergent.

f (x)dx is convergent.

Proof:

By hypothesis (iii), lim

R a

g(x)dx = B say. Let F (R) =

cally increasing and by hypothesis (ii) it is bounded above since

0 F (R) =
Z

Hence, lim

R a

f (x)dx

f (x) exists and is finite, so the integral

f (x)dx. Then F (R) is monotoni0

R
a

R
a

g(x)dx B.
f (x) dx is convergent.

Theorem 2. Assume that

(i) f and g are Riemann integrable in [a, R] for every R > a;
(ii) 0Z f (x) g(x), a x < ;
(iii)
Z

Then

f (x)dx is divergent.

g(x)dx is divergent.

Proof:

As in Theorem 1.
Z
x2

Example 3. Show that

dx is convergent.
1 + x7
2
Z
x2
1
dx
Note that on [2, ), 0 <
< 3/2 . But since
is convergent, so is the given integral, by
7
x
x3/2
x +1
2
the comparison test.
Z
x3

dx is divergent.
Example 4. Show that
1 + x7
2
Observe that

But

x3
1
1
1

, 2 x < .
=
7
7
7
x
x +1
x 1+x
1+2

1
dx is divergent. Hence, by Theorem 2, the given integral is divergent.
x

29

Test for Absolute Convergence.

Z
Z
f (x) dx is said to be absolutely convergent if
Definition. The integral

Definition. The integral

f (x) dx is said to be conditionally convergent if

|f (x)| dx is convergent.
Z

f (x) dx is conver-

gent, but not absolutely convergent.

Z
sin x
Example 5. Show that
dx is absolutely convergent.
x2
1
Note that
| sin x|
1
0
2 , 1 x < .
2
x
x
Z
Z
| sin x|
dx
is
convergent
by
Example
1,
the
integral
dx is convergent by Theorem 1, so the
Since
2
x
x2
1
1
given integral is absolutely convergent.
Theorem 3. Assume that
(i) fZ is Riemann integrable in [a, R] for every R > a;

(ii)

aZ

|f (x)| dx is convergent.

Then

f (x) dx is convergent.

Proof : Since 0 |f (x)| + f (x) 2|f (x)|, on the interval (a, ) the integral is convergent by Theorem 1.
Z
[|f (x)| + f (x)] dx
a

is convergent. If we subtract from this integral the convergent integral in hypothesis (ii), we obtain the
required convergence.
Z
sin x
Example 6. Show that
dx is conditionally convergent.
x

Solution. In the interval k x (k + 1) where k = 1, 2, . . . we have

| sin x|
| sin x|

.
x
(k + 1)

(2)

Hence,
Z
If n R < (n + 1)

(k+1)
k

1
| sin x|
dx
x
(k + 1)
Z

(k+1)

| sin x|dx =

2
.
(k + 1)

n1
2X 1
| sin x|
dx
.
x

k+1

(3)

k=1

| sin x|
dx
x

is divergent. Hence, the integral is not absolutely convergent. However, we show that this integral is
As R , so does n and so does the right hand side of the inequality (3). This proves that

30

convergent. Indeed, integrating by parts, we have

Z R
Z R
sin x
cos x
1 cos R
dx =
+
dx.
(4)
x

R
x2

R x
x
| x12 , by Theorem 1 and Example 1, cos
dx converges absolutely and hence, it converges.
Since | cos
x2
x2
Letting R 7 in (4), we now conclude that the integral converges.
Limit Test for convergence.
Theorem 4. Assume that
(i) f (x) is Riemann integrable in [a, R] for every R > a;
(ii) lim xp f (x) = A for some p > 1.
x

Then

|f (x)| dx is convergent.

(5)

Proof : From hypothesis (ii), we have lim xp |f (x)| = |A|. Hence there is a number b such that
x

Since,

xp |f (x)| |A| + 1,

b < x < .
Z
dx
|f (x)| is convergent.
, p > 1 is convergent, by Theorem 1, we have
xp
a

Limit test for divergence.

Theorem 5. Assume that
(i) f (x) is Riemann integrable in [a, R] for every R > a;
(ii) lim xp f (x) = A 6= 0 (or = ) for some p 1.
x
Z
Then
f (x)dx is divergent.
a

The test fails if A = 0.

Proof : Case I : Let A > 0. Then a number b > 0 exists such that
xp f (x) >

dx
is divergent, by Theorem 2,
xp
b
If A = +, the argument is similar.
Since

A
,
2
Z

b x < .
f (x) dx is divergent whence the desired conclusion.

Case II : A < 0 (or A = ). In this case the integral

[f (x)] dx may be treated by Case I.

Next we show that the test fails when A = 0. Consider the integrals
Z
Z
dx
dx
and
.
2
x
x
log
x
2
1
31

(6)

In each case, for p = 1, we obtain A = 0. Hence, the first integral is convergent and the second integral is
divergent.
Illustration of Theorem 4 and Theorem 5.
Z
2
Example 7. Show that
ex dx is convergent.
0

Note that

lim x2 f (x) = lim x2 ex = 0.

Hence, the result follows.

Z
cos x

dx is absolutely convergent.
Example 8.
1 + x3
0
Observe that
lim x5/4 f (x) = lim

x x1/4

cos x

= 0.
1 + x3

Therefore, the conclusion follows from Theorem 4.

Z
dx

is divergent.
Example 9.
1 + 2x2
0
Note that

x
1
6 0.
lim f (x) = lim
= =
x
x
2
1 + 2x2
Therefore, the result follows from Theorem 5.
Improper Integrals of the Second Kind.
Let f (x) be unbounded at a and let it be Riemann integrable in [a+, b] for all  > 0 such that 0 <  < ba.
Z b
f (x) dx.
Let F () =
a+
Z b
f (x) dx is said to be convergent, if lim F () = A, for some
Definition. The improper integral
0+

A IR. In this case A is called the value of the integral. The integral (7) is said to be divergent, if it is
not convergent.
Likewise, if f (x) is unbounded at b and it is Riemann integrable in [a, b ] for all  > 0 such that
Z b
0 <  < b a, let F () =
f (x) dx.
a

Definition. The improper integral (7) is said to be convergent, if lim F () = A, for some A IR. In
0+

this case again, A is called the value of the integral, and the integral (7) is said to be divergent, if it is
not convergent.
If the function f is unbounded at a point c in (a, b), then the integral (7) is said to be convergent if both
Z c
Z b
the integrals
f (x) dx and
f (x) dx converge according to the above definitions.
a

Remark. In the above definitions and the results to follow, in place of hypothesis of Riemann integrability
of f , one may simply assume f to be continuous in the corresponding interval. For instance, if f is
unbounded at a, we may assume f to be continuous in the interval (a, b].

32

Example 10. Discuss the convergence of

b
a

dx
,
(x a)p

where p is real.

Case I: < p 0, the integral is proper.

Case II: p = 1.
lim

0+

Hence, the integral is divergent.

Case III: 1 < p <
lim

0+

a+

a+

dx
= lim [log(b a) log ] = .
x a 0+


dx
(b a)1p
1p
= lim

(x a)p 0+
1p
1p

does not exist. Hence, the integral is divergent.

Case IV: 0 < p < 1.
Z b
(b a)1p
dx
=
.
lim
1p
0+ a+ (x a)p

Hence, the integral is convergent.

Z b
dx
Thus the integral
is convergent for < p < 1 and divergent for p 1.
(x

a)p
a
Theorem 6.

(Comparison Test) Assume that

(i) f (x) and g(x) are Riemann integrable in [a + , b] for every  > 0 with 0 <  < b a;
(ii) 0 f (x) g(x), a < x b;
Z b
(iii)
g(x) dx is convergent.
a

Then

f (x) dx

is convergent.

Proof : For  > 0,

b
a+

f (x) dx

a+

g(x)dx

b
a

g(x) dx. As  0 the integral on the left is increasing,

but remains bounded above. Consequently, it approaches a limit.

Theorem 7. Assume that
(i) f (x) and g(x) satisfy condition (i) of Theorem 6,
(ii) 0 g(x) f (x), a < x b and that
Z b
(iii)
g(x) dx is divergent
a

Then

f (x) dx is divergent.
a

33

Proof : This is similar, to that of Theorem 2, except for a change in the limits of integration.
Theorem 8. Assume that
(i) f (x) is Riemann integrable in [a + , b] for every  > 0 with 0 <  < b a;
Z b
|f (x)| dx is convergent.
(ii)
a

Then

f (x) dx is convergent.
a

Proof : It is same as in Theorem 3 except for a change in the limits of integration.

Limit Tests.
Theorem 9. Assume
(i) f (x) satisfies condition (i) of Theorem 8 and
(ii) lim (x a)p f (x) = A,
xa+

Then

Rb
a

for some p, 0 < p < 1.

|f (x)| dx is convergent.

Proof : By hypothesis (ii), there is a c such that (x a) p |f (x)| |A| + 1, a < x c < b. Since
Z b
Z b
1
dx is convergent for 0 < p < 1, by Theorem 7, the integral
|f (x)| dx is convergent.
p
a
a (x a)
Theorem 10. Assume that
(i) f (x) satisfies the condition (i) of Theorem 1;
(ii) lim (x a)p f (x) = A 6= 0 (or ), for some p 1.
xa+

Then

f (x) dx

is divergent. The test fails if A = 0

Proof : Proceed exactly as in Theorem 5. To show that the test fails when A = 0, consider the integrals
Z 1
Z 1/2
dx
dx

.
(a)
dx (b)
x
x log( x1 )
0
0
For p = 1, A = 0 in both the integrals, but (a) is convergent and (b) is divergent.
Z 1/2
1
Example 11. Discuss the convergence of
(log ) dx.
x
0

For 0 the integral is proper and, for > 0, lim xf (x) = 0. Hence, by the limit test the integral is
convergent.

x0+

34

Example 12. Discuss the convergence of

tx1 et dt.

Let f (t) = tx1 et . There are four cases to consider.

Case I (x > 1) lim f (t) = lim tx1 et = 0, and hence the integral is proper.
t0+

t0+

Case II (x = 1) lim f (t) = lim et = 1, hence the integral is proper.

t0+

Case III (0 < x < 1) lim t

t0+
1x

f (t) = 1.

t0+

Case IV (x 0) lim tf (t) = 1 6= 0 for x = 0 and the limit is + for x < 0. Hence, the integral is
t0+

convergent for 0 < x < 1, proper for x 1 and divergent for x 0.

Example 13. Show that the Beta Integral
Z 1
xm1 (1 x)n1 dx
B(m, n) =
0

(m)(n)
=
(m + n)

exists for m, n > 0.

The integral is proper for m 1, and n 1. The integrand is unbounded at 0 if m < 1 and it is unbounded
at 1 if n < 1.
Let m < 1 and n < 1. We take any number, say 12 , between 0 and 1 and examine the convergence of the
improper integrals
Z 1/2
Z 1
m1
n1
xm1 (1 x)n1 dx.
x
(1 x)
dx and
1
2

at 0 and 1 respectively.
Convergence at 0:
lim x1m xm1 (1 x)n1 = 1.

x0+

Since
Z 1/2
0

1/2
0

1
x1m

dx is convergent for 1 m < 1 or m > 0, by the comparison test

xm1 (1 x)n1 dx is convergent at x = 0, for m > 0. The convergence of

Z

1
1/2

xm1 (1 x)n1 dx

at 1 is equivalent to the convergence of

Z 0

(1 t)m1 (t)n1 dt
1/2

(using the substitution 1 x = t)

at 0. Since,
Z

1/2

tn1 (1 t)m1 dt

is convergent if n > 0,

the given integral exists if m > 0 and n > 0.

35

Example 14. (The Gamma integral) Show that the gamma integral (a) =
if a > 0.

xa1 ex dx converges

Here the integrand is unbounded at 0, if a < 1. Thus, we have to examine the convergence at as well as
at 0. Consider any positive number, say 1, and examine the convergence of
Z
Z 1
f (x)dx
f (x)dx and
1

where f (x) = xa1 ex at 0 and , respectively.

(i) Convergence at 0: Let a < 1. Note that lim x

1a

x0

or a > 0.

a1 x

= 1 and

1
dx is convergent if 1a < 1
x1a

(ii) Convergence at : We know that

ex > xa+1

for all a.

1
dx is convergent. Hence, by the comparison test,
x2
1
Z
convergent for all a. Thus, the integral
xa1 ex dx is convergent for a > 0.
1
.
x2

Thus xa1 ex <

But

xa1 ex dx is

Example
Z 115. Let us discuss the
Z 1convergence of the following integrals
dx
dx

(i)
(ii)
.
2
1/2
1 x2
0 x +x
0
The integrand in (i) is unbounded at 0 and the one in (ii) is unbounded at 1. Since
lim

0+

1
x1/2

1
x2 +x1/2

dx = lim 2 2  = 2,
0+

the integral in (i) is convergent. Since

Z 1



dx

= lim sin1 (1 ) sin1 0 = sin1 1 = ,

lim
+
2
2
0
0+ 0
1x
the integral in (ii) is convergent.
Example
of the improperZintegral
Z 216. Discuss the convergence
Z

x
dx
dx

.
(i)
dx
(ii)
(iii)
2+ x
1

x
x
1
+
x2
0
0

(i)Let f (x) =

x
1x .

The function f is unbounded at x = 1. We write

Z

Since f (x) =

1
1x

2
0

xdx
=
1x

1
0

xdx
+
1x

xdx
.
1x

1, the integral becomes

Z

f (x)dx =
0

1
0

dx
+
1x
36

2
1

dx
2.
1x

1
x1/2

and

Also,

Z 2
dx
dx
= lim ( ln()) = , lim
= lim ln() = .
1 x 70+
70+ 1+ 1 x
70+
70+ 0
R 2 dx
R2
R 1 dx
Hence, the two integrals 0 1x , 1 1x are both divergent . Therefore, the integral 0 f (x)dx is also
divergent.
lim

1

(ii) Here the integral is an improper integral of the third kind. This can be expressed as a sum of improper
integrals of first kind and second kind.
Z
Z 1
Z
dx
dx
dx
=
+
= I1 + I2 (say)
2
2
2
x + x
x + x
0
0 x + x
1
For 0 x 1, x2 +1x 1x and so by Theorem 9, I1 is convergent. For 1 x, x2 +1x
by Theorem 1 and Example 1, I2 is convergent. Thus the given integral is convergent.

1
,
x2

and again

(iii) We formally write

Z

Now

Likewise,

dx
+
1 + x2

dx
.
1 + x2

Z R


dx
dx
=
lim
= lim tan1 (R) = .
2
2
R7 0 1 + x
R7
1+x
2
0
Z
dx
dx

= . Hence,
is convergent with value equal to .
2
2
1+x
2
1 + x
Z

dx
=
1 + x2

The integral test.

Example 17. Assume that
(i) f (x) is a non-increasing function for x 1,
(ii) f (x) 0 for x 1 and
(iii) f is Riemann integrable in [1, R] for every R > 1.
Then the series

n=1 f (n)

and the integral

R
1

f (x)dx are convergent or divergent together.

Proof : For n IN, by hypothesis 1, if n x n + 1 then f (n) f (x) f (n + 1).Integrating from n to

n + 1, we obtain
Z n+1
Z n+1
Z n+1
f (n) dx
f (x) dx
f (n + 1) dx,
n

which gives
f (n)

n+1

f (x) dx f (n + 1).

Thus, for N IN, We have

N
1
X
n=1

f (n)

f dx

N
1
X
n=1

37

f (n + 1) =

N
X
k=2

f (k).

R
R
P
If 1 f dx converges toP
A, then by the second part of above inequality N

k=2 f (k)
P 1 f dx A, which

implies that the series k=2 f (k) is convergent. On the other hand, if the series n=1 f (n) is convergent
with sum S , then fix R > 1 and pick N IN such that N > R. By the first part of the above inequality,
Z

f dx

This proves that the integral

similar and it is left to reader.

R
1

N
1

f dx

N
1
X
n=1

f dx S.

f (x)dx is convergent. The proof of the statement about divergence is

Example 18. Discuss the convergence of the series

1
, where > 0.
n(ln n)
n=2

Let f (x) = x(ln1x) . Since > 0, f (x) is non-negative and non-increasing for x 2. By the integral test,
R
P
P
1
the series
n=2 n(ln n) =
n=2 f (n) and the integral 2 f dx converge or diverge together. Now
Z

f dx =

1
1
=
(ln x)1 |R
2,

x(ln x)
1

if 6= 1.

RR
RR
2)1
, if
For = 1, 2 f dx = ln(ln x)|R
. Hence, limR7 2 f dx = , if 1 and is equal to = (ln1
R 2
> 1. Thus the integral 2 f dx converges if > 1 and diverges if 0 1. By the integral test, the
given series converges if > 1 and diverges if 0 1.

38