DEPARTMENT OF MATHEMATICS

I.I.T. MADRAS
MA 1101 - Calculus
Solutions to Assignment - 3
(1) Prove the following:
(a) It is not possible that a series converges to a real number ℓ and also diverges to −∞.
(b) It is not possible that a series diverges to ∞ and also to −∞.
P
(a) Suupose aj converges to ℓ and also diverges to −∞. Then we have natural numbers k, m such
that for every n ≥ k, ℓ − 1 < nj=1 aj < ℓ + 1. And also for all n ≥ m, nj=1 aj < ℓ − 2. Choose
P P

M = max{k, m}. Then both inequalities hold for n = M. But this is not possible.
P
(b) Suppose aj diverges to both ∞ and to −∞. Then we have natural numbers k, m such that for
each n ≥ k, j=1 aj > 1 and for each n ≥ m, nj=1 aj < −1. Choose M = max{k, m}. Then both
Pn P

the inequalities hold for n = M. But this is impossible.

(2) Let k be any real number. Prove the following:
P P P P P
(a) If both the series an and bn converge, then the series (an + bn ), (an − bn ) and kan
converge.
P P P P
(b) If an converges and bn diverges to ±∞, then (an + bn ) diverges to ±∞, and (an − bn )
diverges to ∓∞.
P P
(c) If an diverges to ±∞, and k > 0, then kan diverges to ±∞.
P P
(d) If an diverges to ±∞, and k < 0, then kan diverges to ∓∞.
P P
(a) Suppose an converges to ℓ and bn converges to s. Let ϵ > 0. Then we have natural numbers
k, m such that for all n ≥ k, | j=1 aj − ℓ| < ϵ/2; and for all n ≥ m, | nj=1 bj − s| < ϵ/2. Choose
Pn P

M = max{k, m}. Then for all n ≥ M, both the inequalities hold. So, we obtain
n
X n
X n
X
(aj + bj ) − (ℓ + s) ≤ aj − ℓ + bj − s < ϵ/2 + ϵ/2 = ϵ.
j=1 j=1 j=1

Similarly, the other two are proved.

P P
(b) Suppose an converges to ℓ and bn diverges to ∞. Let r > 0. Then, we have natural numbers
k, m such that for all n ≥ k, ℓ − 1 < j=1 aj < ℓ + 1; and for all n ≥ m, nj=1 bj > r + |ℓ| + 1.
Pn P

Choose M = max{k, m}. Then all the three inequalities hold for n ≥ M. But then for all n ≥ M,
n
X n
X
ℓ−1< aj , r + |ℓ| + 1 < bj .
j=1 j=1
Pn
That is, for all n ≥ M, r ≤ ℓ − 1 + r + |ℓ| + 1 < j=1 (aj + bj ).
Similarly, other cases are proved.
P
(c) Suppose an diverges to ±∞, and k > 0. Let r ∈ R. We have m ∈ N such that for all n ≥ m,
Pn Pn
j=1 aj > r/k. Then for all such n, j=1 (kaj ) > r.
Similarly other cases are proved.
P
(d) Suppose an diverges to ∞, and k < 0. Let r ∈ R. We have m ∈ N such that for all n ≥ m,
Pn Pn
j=1 aj > r/k. Then for all such n, j=1 (kaj ) < r, since k < 0.
Similarly other cases are proved.
(3) Give examples for the following:
P P P
(a) an and bn both diverge, but (an + bn ) converges to a nonzero number.
2
P P P
(b) an and (an + bn ) diverges to ∞.
bn both diverge, and
P P P
(c) an and bn both diverge, and (an + bn ) diverges to −∞.
(a) 1 + 1 + 1 + · · · diverges; 2 + (−1) + (−1) + · · · also diverges.
But (1 + 2)) + (1 + (−1)) + · · · = 3 + 0 + · · · converges to 3.

(b) 1 + 2 + 3 + 4 + · · · diverges; −1 − 1 − 1 − 1 − · · · also diverges.

And (1 − 1) + (2 − 1) + (3 − 1) + · · · = 0 + 1 + 2 + 3 + · · · diverges to ∞.

(c) −1 − 2 − 3 − 4 − · · · diverges; 1 + 1 + 1 + 1 + · · · also diverges.

And (−1 + 1) + (−2 + 1) + (−3 + 1) + · · · = 0 − 1 − 2 − 3 − · · · diverges to −∞.
ln(1 + x)
(4) Determine power series expansion of the functions (a) ln(1 + x) (b) .
1−x
1
(a) For −1 < x < 1, = 1 − x + x2 − x3 + x4 − · · ·.
1+x
Integrating term by term and evaluating at x = 0, we obtain
x2 x3 x4
ln(1 + x) = x − + − + ··· for − 1 < x < 1.
2 3 4
(b) Using the results in (a) and the geometric series for 1/(1 − x), we have
∞ ∞
ln(1 + x) X (−1)n−1 xn X n
= · x for − 1 < x < 1
1−x n
n=1 n=0

an xn .
P
For obtaining the product of the two power series, we need to write the first in the form
(Notice that for the second series, each bn = 1.) Here, the first series is
∞
X (−1)n−1
ln(1 + x) = an xn , where a0 = 0 and an = for n ≥ 1.
n
n=0
∞
ln(1 + x) X
Thus the product above is = cn xn , where
1−x
n=0

1 1 (−1)n−1
cn = a0 bn + a1 bn−1 + · · · + an b0 = a0 + a1 + · · · an = 1 − + − ··· + .
2 3 n
1
(5) The function converges for −1 < x < 1. However, prove that it has power series representation
1−x
around any c ̸= 1.
∞
1 1 1 1 X 1
= x−c = (x − c)n .
1−x 1 − c 1 − 1−c 1−c (1 − c)n
n=0

This power series converges for all x with |x − c| < |1 − c|, i.e., for x ∈ (c − |1 − c|, c + |1 − c|).
1
We also see that the function is well defined for each x ̸= 1.
1−x
∞
X (−1)n
(6) Find the sum of the alternating harmonic series .
n
n=0
1
Consider the power series representation of . Integrating term by term.
1+x
∞ ∞
1 X X xn+1
= (−1)n xn ⇒ ln(1 + x) = (−1)n .
1+x n+1
n=0 n=0

This series converges for x = 1. Evaluating it at x = 1, we have

1 1 1 1
ln 2 = 1 − + − + − ···
2 3 4 5
 3
Z a
sin x
(7) Give an approximation scheme for dx where a > 0.
0 x
Using the Maclaurin series for sin x, we have
sin x x2 x4 x6
=1− + − + ···
x 3! 5! 7
Integrating term by term, we get
Z a
sin x a3 a5 a7
dx = a − + − + ···
0 x 3! · 3 5! · 5 7! · 7
Approximations to the integral may be obtained by truncating the series suitably.
1 1 1·3 1 1·3·5 1 π
(8) Show that 1 + + · + · + ··· = .
2 3 2·4 5 2·4·6 7 2
m m(m − 1) 2 m(m − 1)(m − 2) 3
In the binomial series (1 + x) = 1 + mx + x + x + ···
1·2 1·2·3
for |x| < 1, substitute x = −t2 and m = −1/2 to obtain
1 1 1·3 4 1·3·5 6
√ = 1 + t2 + t + t + ···
1−t2 2 2·4 2·4·6
Integrating this power series from 0 to x for any x ∈ (−1, 1), we have
Z x
dt 1 1 1·3 1 5 1·3·5 1 7
sin−1 x = √ = x + · x3 + · x + · x + ···
0 1−t 2 2 3 2·4 5 2·4·6 7
This series also converges for x = 1. It may be seen as follows:
1 · 3 · 5 · · · (2n − 1) 1 2n + 1 2n + 1
Here, leaving the first term, un = · , un+1 = un · · .
2 · 4 · · · 6 · · · (2n)
2n + 1 2n + 2 2n + 3
It follows that for n ≥ 2, un+1 < 4 nun − (n + 1)un+1 .
m
X m−1
X  m−1
X 
Then sm = un = un+1 < 4 (nun − (n + 1)un+1 = 4(u1 − mum ) < 4u1 .
n=2 n=1 n=1
Hence sm is an increasing sequence of positive terms having an upper bound as 4u1 .
∞
X
That is, the series 1 + un converges.
n=1
π
Therefore, for x = 1, the series converges to sin−1 1 = .
2

(9) Find the Fourier series of f (x) given by: f (x) = 0 for −π ≤ x < 0; and f (x) = 1 for 0 ≤ x ≤ π. Say
also how the Fourier series represents f (x). Hence give a series expansion of π/4.

Z π 1 for n = 0
1
an = cos nx dx =
π 0 0 for n ̸= 0


π 2
(−1)n for n odd
 
1 1 − cos nπ 1−
Z
1 
nπ
bn = sin nx dx = = =
π 0 π n nπ 0 for n even
∞
1 2 X sin(2n + 1)x
Hence the Fourier series for f (x) is + .
2 π 2n + 1
n=0
By the convergence theorem for Fourier series, we know that this Fourier series converges to f (x)
for any x ̸= 0. At x = 0, the Fourier series converges to 1/2.

Taking x = π/2, we have

∞ ∞
1 2 X sin(n + 1/2)π 1 2 X (−1)n
1= + = + .
2 π 2n + 1 2 π 2n + 1
n=0 n=0
4

∞
π X (−1)n
Therefore, = .
4 2n + 1
n=0
∞
π2 X 1
(10) Considering the fourier series for |x|, deduce that = .
8 (2n + 1)2
n=0
Cinsider f (x) = |x| in the interval [−π, π]; extended to R with period 2π. Now, it is an even function.
Rπ
Thus each bn is 0. Next, a0 = (2/π) 0 x dx = π. And for n > 0,

2 π 2 (−1)n − 1
Z  
an = x cos nx dx = .
π 0 π n2
−4
That is, a2n = 0, a2n+1 = for n = 1, 2, 3 . . .
π(2n + 1)2
By the convergence theorem for Fourier series, we have
∞
π 4 X cos(2n + 1)x
|x| = − for x ∈ [−π, π].
2 π (2n + 1)2
n=0

∞
π2 X 1
Taking x = 0, we have = .
8 (2n + 1)2
n=0
∞
π X (−1)n
(11) Considering the fourier series for x, deduce that = .
4 2n + 1
n=0
Consider f (x) = x for x ∈ [−π, π]. It is an odd function. Hence in its Fourier series, each an = 0.
For n ≥ 1,
π
2 h cos nx iπ 2 π cos nx 2(−1)n+1
Z Z
2
bn = x sin nx dx = −x + dx = .
π 0 π n 0 π 0 n n
∞
X (−1)n+1
Thus the Fourier series for f (x) = x in [−π, π] is 2 sin nx.
n
n=1
∞ ∞
π X (−1)n+1 nπ (−1)n
X
Taking x = π/2, we have = sin = .
4 n 2 2n + 1
n=1 n=0
(12) Considering the fourier series for f (x) given by: f (x) = −1, for −π ≤ x < 0 and f (x) = 1 for
∞
π X (−1)n
0 ≤ x ≤ π. Deduce that = .
4 2n + 1
n=0
Here, f (x) is an odd function. Thus in its Fourier series, each an is 0. For n ≥ 1,
2 π
Z
2 2
bn = sin nx dx = (1 − cos nπ) = (1 − (−1)n ).
π 0 nπ nπ
∞
4 X sin(2n + 1)x
Due to the convergence theorem, we conclude that for x ̸= 0, f (x) = .
π 2n + 1
n=0
Taking x = π/2, we obtain the desired expression for π/4.
(13) Considering f (x) = x2 , show that for each x ∈ [0, π],
∞ ∞
π2 X (−1)n cos nx X n2 π 2 (−1)n+1 + 2(−1)n − 2
+2 = sin nx.
6 n2 n3 π
n=1 n=1

We determine sine and cosine series expansions of f (x) = x2 for 0 ≤ x ≤ π.

The odd and even expansions of f (x) are

−x2 for − π ≤ x < 0
fodd (x) = feven (x) = x2 for − π ≤ x ≤ π.
 x2 for 0 ≤ x < π,
 5

∞
π2 X (−1)n cos nx
We see that, as earlier, feven (x) has the Fourier expansion +4 for x ∈ [0, π].
3 n2
n=1
Due to the convergence theorem of Fourier series, this series sums to x2 in [0, π].

For the sine series expansion, we determine the Fourier series of fodd (x). Here, each an is 0. And for
n ≥ 1,
2 π 2 (−1)n+1 4 (−1)n − 1
Z  
bn = x sin nx dx = 2π + .
π 0 n π n3
Due to the convergence theorem of Fourier series, x2 = ∞
P
n=1 bn sin nx for x ∈ [0, π].
Equating both the sine and the cosine series for f (x) = x2 in [0, π], we obtain the required result.
(14) Represent the function f (x) = 1 − |x| for −1 ≤ x ≤ 1 as a cosine series.
∞
a0 X
It is an even function. Thus its Fourier series is + an cos nπx, where
2
n=1
R1
a0 = 2 0 (1 − x) dx = 1. 
Z 1 Z 1 0 for n even
For n ≥ 1, an = (1 − |x|) cos nπx dx = 2 (1 − x) cos nπx dx =
−1 0 4/(n2 π 2 ) for n odd.
∞
1 4 X cos(2n + 1)πx
Thereofore, 1 − |x| = + 2 for −1 ≤ x ≤ 1.
2 π (2n + 1)2
n=0