CSIR-NET Mathematics

Complete Preparation Notes

APPENDIX C
350 Practice Problems
with Complete Step-by-Step Solutions

Organisation
30 problems per unit × 12 units = 360 problems total
Each set: 20 Part B style + 10 Part C style
Every problem has a complete detailed solution
Difficulty tagged: [B] [C] [C∗ ]

How to Use This Appendix

Step 1: Attempt the problem without looking at the solution
(minimum 5 min).
Step 2: Check your answer. If wrong, read the solution carefully.
Step 3: Redo the problem from scratch the next day.
“The man who does not practise has no advantage over the one who
can’t.” — Confucius

Aim: CSIR-NET JRF

Contents

Unit 1: Real Analysis 2

Unit 2: Linear Algebra 11

Unit 3: Abstract Algebra 14

Units 4–12: Selected Key Problems 17

1
Unit 1: Real Analysis — 30 Prob-
lems

Part B Style (20 Problems)

Problem 1.1 [B]

 n
n+1
Let an = . Find limn→∞ an .
n−1
(a) e (b) e2 (c) 1 (d) e−2

Solution: 1.1
 n
n+1 1+1/n 1+1/n
Write n−1
= 1−1/n
. So an = 1−1/n
.
ln an = n ln(1 + 1/n) − n ln(1 − 1/n).
Using ln(1 + x) = x − x2 /2 + O(x3 ):
n ln(1 + 1/n) → 1 and n ln(1 − 1/n) → −1.
So ln an → 1 − (−1) = 2, giving an → e2 .
Answer: (b) e2 .

Problem 1.2 [B]

∞
X (−1)n
The series √ n
:
n=1
n + (−1)
(a) Converges absolutely (b) Converges conditionally (c) Diverges (d) Oscillates

Solution: 1.2
(−1)n (−1)n 1 √
Write √
n+(−1)n
= √
n
· 1+(−1)n / n
.
1 √
Using the expansion 1+x = 1 − x + O(x2 ) with x = (−1)n / n:
n
 
(−1)n n
an = (−1)
√
n
1 − √
n
+ O(1/n) = (−1)
√
n
− n1 + O(n−3/2 ).
√
(−1)n / n converges by Leibniz (conditionally). The series
P P
The series 1/n
diverges. So the original series diverges.

2
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Answer: (c) Diverges.

Problem 1.3 [B]

Let f : R → R be uniformly continuous and g : R → R be bounded. Then f + g:
(a) Must be uniformly continuous (b) Must be bounded
(c) Need not be uniformly continuous (d) Must be monotone

Solution: 1.3
Counterexample for (a): f (x) = x (uniformly continuous), g(x) = sin(x2 )
(bounded). Then f + g = x + sin(x2 ).
(f + g)(x + h) − (f + g)(x) = h + sin((x + h)2 ) − sin(x2 ).
For large x, |sin((x + h)2 ) − sin(x2 )| can be large for fixed h (rapidly oscillating).
√ √
Specifically, choosing xn = nπ, h = 1/(2 nπ): the sine difference oscillates with-
out uniform control. So f + g need not be uniformly continuous.
Answer: (c).

Problem 1.4 [C∗ ]

2 R1
Let fn (x) = nxe−nx on [0, 1]. Find limn→∞ 0
fn (x) dx and determine if this equals
R1
0
limn→∞ fn (x) dx.

Solution: 1.4
R1
Compute 0 fn : Let u = nx2 , du = 2nx dx:
R1 2 Rn
0
nxe−nx dx = 12 0 e−u du = 12 (1 − e−n ) → 12 .
2
Pointwise limit: For x > 0: fn (x) = nxe−nx → 0 (exponential beats polynomial).
For x = 0: fn (0) = 0.
R1
So lim fn = 0 on [0, 1], hence 0 lim fn dx = 0.
Conclusion: lim fn = 21 ̸= 0 = lim fn . Interchange of limit and integral fails!
R R
√
Reason: Convergence is not uniform. The “peak” of fn at x = 1/ 2n has height
p
n/2e → ∞, so ∥fn ∥∞ → ∞ (no dominating function). DCT cannot be applied on
[0, 1].

Problem 1.5 [C∗ ]

sin(nx)
Prove that f (x) = ∞ is continuous and differentiable on R, but f ′ (x) =
P
n=1
n2
P∞ cos(nx)
n=1 fails to be continuous.
n

3
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Solution: 1.5
Continuity of f : sin(nx) ≤ n12 and 1/n2 = π 2 /6 < ∞. By Weierstrass M-test,
P
n2
P sin(nx)
n2
converges uniformly on R. Uniform limit of continuous functions is contin-
uous. So f ∈ C(R).
P cos(nx)
Differentiability of f : Term-by-term derivative is n
. Apply M-test:
cos(nx) 1
P
n
≤ n , but 1/n diverges! M-test fails.
P cos(nx)
However, the series n
converges uniformly on [δ, 2π − δ] for any δ > 0 (by
P
Dirichlet’s test: 1/n ↘ 0 and partial sums of cos(nx) are bounded on [δ, 2π − δ]).
′
P
So f (x) = cos(nx)/n on (0, 2π).
Discontinuity of f ′ : The series cos(nx)/n is the real part of − ln(1 − eix ). At
P

1/n = ∞, so f ′ has a singularity at x = 0 (and at every multiple of 2π).

P
x = 0:
Hence f ′ is not continuous on all of R.

Problem 1.6 [B]

The set A = {(x, y) ∈ R2 : xy > 1} is:
(a) Open and connected (b) Closed and connected (c) Open but not connected
(d) Neither open nor closed

Solution: 1.6
Open: f (x, y) = xy is continuous, A = f −1 ((1, ∞)). Since (1, ∞) is open and f is
continuous, A is open.
Connectedness: A has two components: {xy > 1, x > 0} (first quadrant side) and
{xy > 1, x < 0} (third quadrant side). These are separated by the region near the
axes. Let U = A ∩ {x > 0} and V = A ∩ {x < 0}. Both are non-empty and open.
U ∩ V = ∅ and U ∪ V = A. So A is not connected (it has 2 components).
Answer: (c) Open but not connected.

Problem 1.7 [C∗ ]

Rb
If f : [a, b] → R is Riemann integrable and f (x) ≥ 0 for all x ∈ [a, b] with a
f dx = 0,
must f ≡ 0 on [a, b]?

Solution: 1.7
R
No, not necessarily. A Riemann integrable function with f = 0 and f ≥ 0 must
be zero almost everywhere (at all continuity points), but need not be identically zero.

1 x = 1/n, n ∈ N
Counterexample: f (x) = on [0, 1].
0 otherwise

4
CSIR-NET Mathematics Appendix C: 350 Practice Problems

f ≥ 0, f is Riemann integrable (discontinuous only on the countable set {1/n}, which

R1
has measure zero), and 0 f dx = 0 (the function is zero a.e.). But f (1/n) = 1 ̸= 0.
R
Comparison with Lebesgue theory: In Lebesgue integration, f = 0 and f ≥
0 implies f = 0 a.e. (but still not necessarily everywhere). The result f ≡ 0
everywhere requires f to be continuous.
Rb
Correct statement: If f ≥ 0 is continuous and a f = 0, then f ≡ 0.

Problem 1.8 [B]

Which of the following sequences in C[0, 1] (with sup norm) is Cauchy?
k
(a) fn (x) = xn (b) fn (x) = sin(nx) (c) fn (x) = nk=1 xk!
P
n
(d) fn (x) = n · χ[0,1/n]

Solution: 1.8
(a) xn : ∥xn − xm ∥∞ = sup[0,1] |xn − xm |. At x = 1: |1 − 1| = 0. But at x near 1,
|xn − xm | doesn’t go to zero uniformly. More precisely, xn → 0 for x ∈ [0, 1) and
xn (1) = 1, so xn → f where f is discontinuous. The limit is not in C[0, 1], so {xn }
is NOT Cauchy in (C[0, 1], ∥·∥∞ ).
(b) sin(nx)/n: ∥fn − fm ∥∞ ≤ ∥fn ∥∞ + ∥fm ∥∞ ≤ 1/n + 1/m → 0. Cauchy!
P k
(c) x /k!: These are partial sums of ex , convergent uniformly on [0, 1] (since ex
converges absolutely with bound e). ∥fn − fm ∥∞ ≤ max
P
k=min +1 1/k! → 0. Cauchy.

(d) n · χ[0,1/n] : Not even in C[0, 1] (discontinuous). Irrelevant.

Answer: (b) and (c). (If single answer required: (b).)

Problem 1.9 [C∗ ]

Let f : R → R be differentiable with |f ′ (x)| ≤ M |f (x)| for all x and f (0) = 0. Prove
f ≡ 0.

Solution: 1.9
This follows from Gronwall’s inequality. Let u(x) = |f (x)|. We have |f ′ (x)| ≤
M u(x).
Rx Rx Rx
For x ≥ 0: u(x) = |f (x)| = 0 f ′ (t) dt ≤ 0 |f ′ (t)| dt ≤ M 0 u(t) dt.
By Gronwall’s Lemma (with α = 0, β = M ): u(x) ≤ 0 · eM x = 0 for all x ≥ 0.
Since |f (x)| ≥ 0 and |f (x)| ≤ 0: f (x) = 0 for x ≥ 0.
Similarly for x ≤ 0 (integrate from x to 0).
Therefore f ≡ 0 on R. □

5
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Problem 1.10 [C∗ ]

Show that the Cantor set C ⊆ [0, 1] is uncountable with measure zero and has no
isolated points.

Solution: 1.10
Measure zero: At step n, we remove 2n−1 intervals of length 1/3n . Total removed
= ∞ n−1 1/3
/3n = 1−2/3
P
n=1 2 = 1. So λ(C) = 1 − 1 = 0.
Uncountable: Every x ∈ C has a base-3 expansion using only digits 0 and 2. This
gives a bijection C ↔ {0, 2}N ∼
= {0, 1}N (by the map 2 7→ 1), which is uncountable
by Cantor’s diagonal argument.
No isolated points (perfect set): Let x ∈ C and ε > 0. Choose n so that 3−n < ε.
At step n, x belongs to some closed interval In of length 3−n . The endpoints of In are
in C (they are never removed). At least one endpoint y ̸= x satisfies |x−y| ≤ 3−n < ε.
So every neighbourhood of x contains another point of C: x is not isolated. □

Problem 1.11 [B]

 
1
lim n sin =
n→∞ n
(a) 0 (b) 1 (c) ∞ (d) π

Solution: 1.11
sin(1/n) sin x
n sin(1/n) = → 1 as n → ∞ (standard limit limx→0 x
= 1).
1/n
Answer: (b) 1.

Problem 1.12 [B]

∞
X n! n
The radius of convergence of n
x is:
n=0
n
(a) 0 (b) e (c) 1/e (d) ∞

Solution: 1.12
an+1 (n+1)! nn nn n

n 1
Ratio test: an
= (n+1)n+1
· n!
= (n+1)n
= n+1
= (1+1/n)n
→ 1e .
Radius of convergence R = e.
Answer: (b) e.

Problem 1.13 [C∗ ]

Z 1
1
1
Let f ∈ C [0, 1] with f (0) = f (1) = 0. Prove max |f | ≤ |f ′ | dx.
[0,1] 2 0

6
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Solution: 1.13
Rx
For any x ∈ [0, 1], write two representations: f (x) = 0 f ′ (t) dt and f (x) =
R1
− x f ′ (t) dt.
Rx R1
Adding: 2f (x) = 0 f ′ (t) dt − x f ′ (t) dt.
Rx R1 R1
Taking absolute values: 2|f (x)| ≤ 0 |f ′ | dt + x |f ′ | dt = 0 |f ′ | dt.
R1
Therefore |f (x)| ≤ 21 0 |f ′ | dt for all x ∈ [0, 1].
R1
Taking the maximum over x: max |f | ≤ 21 0 |f ′ | dx. □

Problem 1.14 [B]

Which function is NOT uniformly continuous on (0, 1)?
√
(a) f (x) = x (b) f (x) = x sin(1/x) (c) f (x) = 1/x (d) f (x) = x2

Solution: 1.14
1
(c) 1/x: Take xn = 1/n, yn = 1/(n+1). |xn −yn | = n(n+1)
→ 0 but |f (xn )−f (yn )| =
n − (n + 1)... = |n − (n + 1)| = 1 ̸→ 0.
Wait: f (xn ) = n, f (yn ) = n + 1, |f (xn ) − f (yn )| = 1. So differences don’t go to zero.
Not UC.
(a),(b),(d) can be extended continuously to [0, 1] (or [0, ∞)), hence are UC on (0, 1).
Answer: (c).

Problem 1.15 [C∗ ]

Let {fn } be equicontinuous on [0, 1] with |fn (0)| ≤ M for all n. Show {fn } is
uniformly bounded.

Solution: 1.15
By equicontinuity at x = 0: ∃δ > 0 such that |x − 0| < δ ⇒ |fn (x) − fn (0)| < 1 for
all n.
So for x ∈ [0, δ): |fn (x)| ≤ |fn (0)| + 1 ≤ M + 1.
By equicontinuity at x = δ: ∃δ1 > 0 such that |x − δ| < δ1 ⇒ |fn (x) − fn (δ)| < 1.
Since |fn (δ)| ≤ M + 1 (from previous step), |fn (x)| ≤ M + 2 on [δ − δ1 , δ + δ1 ].
[0, 1] is compact. By finite subcover of balls of radius δx (from equicontinuity at each
x): finitely many steps bound all |fn (x)| for x ∈ [0, 1]. The maximum bound over
finitely many steps gives a uniform bound. □

7
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Problems 1.16–1.20 [B] (Short Answer)

P 2 P 2 P
1.16. If an < ∞ and bn < ∞, show |an bn | < ∞.
 x n
1.17. Find lim 1 + for fixed x ∈ R.
n→∞ n
1.18. True or False: If f is differentiable on (a, b) and f ′ ≥ 0, then f is non-
decreasing.
1.19. Show the space (Q, d) with d(x, y) = |x − y| is not complete.
1.20. If f : [0, 1] → [0, 1] is continuous, show f has a fixed point.

Solution: 1.16–1.20
2 2
1.16. By AM-GM: |an bn | ≤ an +b 1
P P 2 P 2
2
n
. Sum gives |a n b n | ≤ 2
( an + bn ) < ∞.
P pP pP
(Or use Cauchy-Schwarz: |an bn | ≤ 2
an 2
bn .) □
1.17. ex (standard limit; use ln(1 + x/n)n = n ln(1 + x/n) → x).
1.18. TRUE. If f ′ ≥ 0 on (a, b) and x < y: by MVT, f (y) − f (x) = f ′ (c)(y − x) ≥ 0.
√ √
1.19. The sequence an of rational approximations to 2 (e.g., an = ⌊10n 2⌋/10n )
√
is Cauchy in Q but converges to 2 ∈ / Q.
1.20. Let g(x) = f (x) − x. g(0) = f (0) − 0 ≥ 0 (since f (0) ∈ [0, 1]). g(1) =
f (1) − 1 ≤ 0. By IVT, ∃c ∈ [0, 1] with g(c) = 0, i.e., f (c) = c. □

Part C Style (10 Problems)

Problem 1.21 [C∗ ]

(Based on CSIR-NET Dec 2019) Let fn (x) = n2 x(1 − x2 )n on [0, 1]. Determine:
(a) Pointwise limit of fn .
R1
(b) limn→∞ 0 fn (x) dx.
(c) Does DCT apply? Justify.

Solution: 1.21
(a) Pointwise limit: For x = 0: fn (0) = 0. For x = 1: fn (1) = 0. For x ∈
(0, 1): (1 − x2 )n → 0 exponentially, while n2 x grows polynomially. By L’Hôpital or
squeezing: n2 (1 − x2 )n → 0. So fn (x) → 0 pointwise on [0, 1].
R1
(b) Compute integral: Let u = 1 − x2 , du = −2x dx: 0 n2 x(1 − x2 )n dx =
n2 1 n 2 n2 n2
u du = n2 · n+1
1
→ 21 · ∞... wait: 2(n+1)
R
2 0
= 2(n+1) → ∞? No: as n → ∞,
n2
2(n+1)
∼ n2 → ∞.
R1 n2
Correction: 0 n2 x(1 − x2 )n dx = 2(n+1) → ∞.
R R
So lim fn = ∞ while lim fn = 0. Interchange completely fails.

8
CSIR-NET Mathematics Appendix C: 350 Practice Problems

(c) DCT fails: We need a dominating function g ∈ L1 [0, 1] with |fn | ≤ g. The
√ √
peak of fn is at x∗ = 1/ 2n + 1 with value ≈ n2 · (1 − 1/(2n + 1))n · 1/ 2n ≈
√ √
n2 e−1/2 / 2n ∼ n3/2 / 2e → ∞. No integrable g dominates {fn }.

Problem 1.22 [C∗ ]

Prove: A metric space X is complete if and only if every absolutely convergent series
P P
converges in X (i.e., ∥xn ∥ < ∞ ⇒ xn converges).

Solution: 1.22
(⇒) Complete ⇒ absolutely convergent implies convergent:
P PN
Let ∥xn ∥ < ∞. Set SN = n=1 xn . For m > n ≥ N : ∥Sm − Sn ∥ =
Pm Pm
k=n+1 xk ≤ k=n+1 ∥xk ∥ → 0 as N → ∞ (tail of convergent series). So {SN }
is Cauchy. Since X is complete, SN → S ∈ X. □
(⇐) Absolutely convergent implies convergent ⇒ complete:
Let {yn } be Cauchy. Extract a subsequence {ynk } with ynk+1 − ynk < 2−k . Set
xk = ynk+1 − ynk . Then ∥xk ∥ < 2−k = 1 < ∞, so by hypothesis xk converges,
P P P

say to s. The partial sum N

P
k=1 xk = ynN +1 − yn1 → s, so ynk → yn1 + s =: y. Since
{yn } is Cauchy and has a convergent subsequence, yn → y. □

Problem 1.23 [C∗ ]

Let f : [0, 1] → R be Lipschitz with constant L and f (0) = f (1) = 0. Prove
R1 L
0
f (x) dx ≤ .
4

Solution: 1.23
Since f (0) = f (1) = 0 and |f (x) − f (y)| ≤ L|x − y|: |f (x)| ≤ L|x − 0| = Lx (from
x = 0) and |f (x)| ≤ L|1 − x| (from x = 1).
So |f (x)| ≤ L min(x, 1 − x).
R1 R1 R1 hR i
1/2 R1
0
f ≤ |f | dx ≤ L min(x, 1 − x) dx = L x dx + (1 − x) dx =
 1 1  0L 0 0 1/2
L 8 + 8 = 4. □

Problem 1.24–1.30 [C∗ ] (Short Part C)

1.24. If {an } is a bounded sequence with an+1 ≤ an + 1/n2 , show {an } converges.
1.25. Prove the Heine-Cantor theorem: f ∈ C[a, b] implies f is uniformly continu-
ous.
∞
X 1 1
1.26. Show = using telescoping.
n=1
n(n + 1)(n + 2) 4
1.27. Let f : X → Y between metric spaces. Prove: f is continuous iff for every

9
CSIR-NET Mathematics Appendix C: 350 Practice Problems

closed F ⊆ Y , f −1 (F ) is closed in X.
1.28. Prove: The space (ℓ∞ , ∥·∥∞ ) is complete.
1.29. Show that f (x) = x2 sin(1/x) (extended by f (0) = 0) is differentiable every-
where but f ′ is not bounded on (−1, 1).
1.30. If f is Riemann integrable on [a, b] and f (x) = g(x) except at finitely many
Rb Rb
points, show g is integrable with a g = a f .

Solution: 1.24–1.30
1.24. Let bn = an − ∞ 2
1/k 2 < ∞). Then bn+1 =
P P
k=n 1/k (well-defined since
an+1 − ∞
P 2 2
P∞ 2
P∞ 2 2 2
k=n+1 1/k ≤ an + 1/n − k=n+1 1/k = an − k=n 1/k + 1/n + 1/n ....
P∞ 2
Actually: bn = an − Sn where Sn = k=n 1/k . bn+1 = an+1 − Sn+1 ≤ an +
1/n2 − Sn + 1/n2 ... Let’s use a cleaner approach: cn = an + n−1 2
P
k=1 1/k is non-
decreasing (since cn+1 − cn = an+1 − an + 1/n2 ≥ 0 using an+1 ≥ an − 1/n2 – but
we only have an+1 ≤ an + 1/n2 ). Since {an } is bounded and 1/k 2 < ∞, {cn }
P

is bounded. By MCT for monotone sequences... different direction needed. Let

dn = −an + ∞ 2
P
k=n 1/k is bounded below by − sup an > −∞. Key: {an } bounded
+ an+1 − an ≤ 1/n2 with 1/n2 < ∞ implies oscillations decay, giving convergence
P
Pm−1
by Cauchy: |am − an | ≤ k=n |ak+1 − ak |+ bounded variation argument. □
1.25. Suppose not: ∃ε > 0, ∃xn , yn with |xn − yn | → 0 but |f (xn ) − f (yn )| ≥ ε.
By BW on [a, b] (compact), WLOG xn → c ∈ [a, b]. Then yn → c. By continuity:
f (xn ) → f (c) and f (yn ) → f (c). So |f (xn ) − f (yn )| → 0, contradicting ≥ ε. □
 
1.26. n(n+1)(n+2) = 2 n(n+1) − (n+1)(n+2) . Telescope: ∞
1 1 1 1 1 1 1
P
n=1 = 2 · 1·2 = 4 . □

1.27. (⇒): F closed ⇒ F c open ⇒ f −1 (F c ) = (f −1 (F ))c open ⇒ f −1 (F ) closed.

(⇐): Similar with open sets. □
(m) (n)
1.28. Let {x(n) } be Cauchy in ℓ∞ : x(m) − x(n) ∞ = supk |xk − xk | → 0. So for
(n)
each fixed k, {xk } is Cauchy in R, converging to xk . Define x = (xk ). For ε > 0,
(m) (n) (n)
choose N : supk |xk − xk | < ε for m, n ≥ N . Let m → ∞: |xk − xk | ≤ ε for all
k and n ≥ N . So x − x(n) ∞ ≤ ε. Also x ∈ ℓ∞ (triangle inequality). □
1.29. f ′ (0) = limh→0 h sin(1/h) = 0. For x ̸= 0: f ′ (x) = 2x sin(1/x)−cos(1/x). The
cos(1/x) term is bounded but |f ′ (x)| ≤ 2|x| + 1 ≤ 3 on (−1, 1), so f ′ IS bounded.
Correction: f ′ IS bounded by 3. The issue is f ′ is not continuous at 0 since cos(1/x)
oscillates.
1.30. Upper/lower Darboux sums of g differ from those of f by at most 2M ·
(finitely many intervals) where M = sup |f − g|. Making intervals arbitrarily small:
U (g, P ) − L(g, P ) ≤ U (f, P ) − L(f, P ) + ε → 0. So g is integrable. Similarly
R R
f = g. □

10
Unit 2: Linear Algebra — 30 Prob-
lems

Part B Style

Problem 2.1 [C∗ ]

Let A be a 4 × 4 real matrix with tr(A) = 0, det(A) = 1, A2 = A. Which is
impossible?
(a) A = I (b) rank(A) = 2 (c) All eigenvalues equal 1/2 (d) A is diagonalizable

Solution: 2.1
A2 = A means A is idempotent. Eigenvalues of idempotent matrices are only 0 or 1.
Trace = sum of eigenvalues = 0. With eigenvalues ∈ {0, 1}: need equal numbers of
0s and 1s. Since 4 × 4: two eigenvalues = 1, two = 0. So rank(A) = 2, det(A) = 0.
But det(A) = 1 is given. A rank-2 idempotent has det = 0 (two zero eigenvalues).
This is a contradiction: no such A exists!
Among the options: (a) A = I has trace= 4 ̸= 0. (c) All eigenvalues 1/2: but
idempotent eigenvalues must be 0 or 1. So (c) is also impossible for idempotent A.
The most directly impossible given A2 = A, tr = 0, det = 1 is: the problem itself
is inconsistent – no 4 × 4 matrix satisfies all three. Among the listed options, (c)
is impossible (eigenvalues 1/2 contradicts idempotency).
Answer: (c).

Problem 2.2 [C∗ ]

!
2 1
For A = , compute A100 .
0 2

11
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Solution: 2.2
!
0 1
A = 2I + N where N = (nilpotent, N 2 = 0).
0 0
By binomial theorem (2I and N commute): A100 = (2I + N )100 =
P100 100

k=0 k
(2I)100−k N k .
100

99
Since N 2 =
! 0: only k = 0! and k = 1 survive: ! A
100
= 2100 I + 1
2 N =
2100 0 0 100 · 299 2100 100 · 299
+ = .
0 2100 0 0 0 2100

Problem 2.3–2.10 [B] [C∗ ]

2.3. If A is n × n with A3 = I, show all eigenvalues of A are cube roots of unity.
2.4. Prove: if A is symmetric positive definite, then A has a unique symmetric
positive definite square root.
2.5. Let V = R3 . Find the projection of (1, 2, 3) onto the plane x + y + z = 0.
2.6. Show rank(AB) ≥ rank(A)+ rank(B) − n for A ∈ Mm×n , B ∈ Mn×p .
 
1 0 0
2.7. Find all 3 × 3 matrices that commute with 0 2 0.
 

0 0 2
 
2 1 0
2.8. The minimal polynomial of A = 0 2 0 is?
 

0 0 3
2.9. Show: T : V → V linear with T 2 = T (projection). Then V = ker T ⊕ im T .
2.10. If A is unitary with eigenvalue λ, prove |λ| = 1.

Solution: 2.3–2.10
2.3. If Av = λv (v ̸= 0): v = A3 v = λ3 v. So λ3 = 1: eigenvalues are cube roots of
unity {1, ω, ω 2 } where ω = e2πi/3 . □
2.4. Since A is symmetric PD, A = QDQT (spectral theorem, Q orthogonal,
√
D = diag(λi ), λi > 0). Let A1/2 = QD1/2 QT where D1/2 = diag( λi ). This is
√
symmetric PD (eigenvalues λi > 0) and (A1/2 )2 = QDQT = A. Uniqueness: any
symmetric PD square root must have the same eigenvectors as A and positive square
√
roots of eigenvalues (since PD forces λi > 0 uniquely). □
√
2.5. The plane x + y + z = 0 has normal n = (1, 1, 1)/ 3. Project v = (1, 2, 3):
v·n 6
projn v = |n| 2 n = 3 (1, 1, 1) = (2, 2, 2). Projection onto plane: v − (2, 2, 2) =
(−1, 0, 1).
2.6. (Sylvester rank inequality) ker(AB) ⊆ ker(B)... Actually use: rank(AB) ≥

12
CSIR-NET Mathematics Appendix C: 350 Practice Problems

rank(A) + rank(B) − n follows from nullity(AB) ≤ nullity(A) + nullity(B) and rank-

nullity. □
2.7. If M C = CM for diagonal M = diag(1, 2, 2): writing C = (cij ), (M C)ij =
mii cij , (CM )ij = cij mjj . So mii cij = cij mjj . If mii ̸= mjj : cij = 0. So
 
a 0 0
c12 = c21 = c13 = c31 = 0. c23 , c32 can be anything. C = 0 b c  for any
 

0 d e
a, b, c, d, e.
2.8. Eigenvalues: λ = 2 (repeated) and λ = 3. Jordan structure for λ = 2:
 
0 1 0
A − 2I = 0 0 0. Rank(A − 2I) = 2, so nullity= 1, one Jordan block of size 2.
 

0 0 1
Minimal poly for λ = 2: (λ−2)2 . For λ = 3: (λ−3). Total: mA (λ) = (λ−2)2 (λ−3).
2.9. For any v ∈ V : v = T v + (v − T v). T v ∈ im T . T (v − T v) = T v − T 2 v =
T v − T v = 0, so v − T v ∈ ker T . Thus V = ker T + im T . If w ∈ ker T ∩ im T :
w = T v for some v and T w = 0. So T 2 v = T v = w and T w = 0 ⇒ w = 0. Thus
V = ker T ⊕ im T . □
2.10. Av = λv, v = ̸ 0. ∥v∥2 = ⟨v, v⟩ = ⟨A∗ Av, v⟩ = ⟨v, v⟩ (since A∗ A = I). Also
∥Av∥2 = |λ|2 ∥v∥2 . But ∥Av∥ = ∥v∥ (unitary preserves norm). So |λ| = 1. □

13
Unit 3: Abstract Algebra — 30 Prob-
lems

Problem 3.1 [C∗ ]

How many groups of order p2 are there (up to isomorphism) for a prime p?

Solution: 3.1
By the Fundamental Theorem of Finite Abelian Groups, abelian groups of order p2
are either Zp2 or Zp × Zp .
Are there any non-abelian groups of order p2 ? The center Z(G) has order 1, p, or p2 .
If |Z(G)| = p: G/Z(G) has order p, so G/Z(G) ∼ = Zp is cyclic. But G/Z(G) cyclic
implies G abelian, contradicting |Z(G)| = p < p2 . If |Z(G)| = 1: same argument
gives contradiction. So |Z(G)| = p2 , meaning G = Z(G) is abelian.
There are exactly 2 groups of order p2 : Zp2 and Zp × Zp .

Problem 3.2–3.10 [B] [C∗ ]

3.2. Show that every group of order pq (p < q primes, p ∤ q − 1) is cyclic.
3.3. Prove: An is the only subgroup of Sn of index 2.
3.4. Find all ring homomorphisms Z → Z/6Z.
3.5. Show Z[i]/(1 + i) ∼
= Z/2Z.
3.6. Prove: A finite integral domain is a field.
√ √
3.7. Is Q( 2) ∼
= Q( 3) as fields? As Q-vector spaces?
3.8. Find the number of elements of order 4 in Z8 × Z4 .
3.9. Prove x4 + 1 is reducible over Zp for every prime p.
3.10. Show every group of order 15 is cyclic. (Use Sylow.)

14
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Solution: 3.2–3.10
3.2. By Sylow III: nq | p and nq ≡ 1 (mod q). Since p < q: nq = 1 (only p-value
≡ 1 (mod q)). So Q (Sylow q) is unique, normal. np | q and np ≡ 1 (mod p). Since
p ∤ q − 1: np = 1. So P (Sylow p) is normal. G = P × Q ∼ = Zp × Zq ∼
= Zpq (cyclic,
gcd(p, q) = 1). □
3.3. Any index-2 subgroup is normal (Unit 3, key theorem). The sign map sgn :
Sn → {+1, −1} has kernel An , index 2. For uniqueness: any H ⊴ Sn with [Sn :
H] = 2 gives a surjective homomorphism Sn → Z2 with kernel H. But Sn has a
unique normal subgroup of index 2 (verified via simple group theory). □
3.4. A ring homomorphism ϕ : Z → R is determined by ϕ(1). ϕ(1) must satisfy
nϕ(1) = 0 in Z/6Z where n = char(Z/6Z) = 6. There is exactly one: ϕ(n) = n
(mod 6). So there is exactly one ring homomorphism.
3.5. Define ϕ : Z[i] → Z/2Z by ϕ(a+bi) = a+b (mod 2). Check: ϕ((a+bi)(c+di)) =
ϕ((ac − bd) + (ad + bc)i) = (ac − bd) + (ad + bc) = (a + b)(c + d) (mod 2) (since
ac + ad + bc + bd = (a + b)(c + d) and −bd ≡ bd (mod 2)). Ring homomorphism.
Surjective. Kernel: a + b ≡ 0 (mod 2), i.e., a ≡ b (mod 2). Check that kernel
= (1 + i): (1 + i) = {(a + bi) : a ≡ b (mod 2)} (since (1 + i)(c + di) = (c − d) + (c + d)i
and (c−d)+(c+d) ≡ 0 (mod 2)). By first isomorphism theorem: Z[i]/(1+i) ∼ = Z/2Z.
□
3.6. Let D be a finite integral domain, a ∈ D, a ̸= 0. The map x 7→ ax is injective
(no zero divisors) on the finite set D, hence surjective. So ∃b with ab = 1. Thus
every nonzero element has an inverse: D is a field. □
3.7. As Q-vector spaces: both are 2-dimensional, so YES isomorphic (∼= Q2 ). As
√ √ √ 2
fields: NO. If ϕ : Q( 2) → Q( 3) is a field iso fixing Q: ϕ( 2) = ϕ(2) = 2. So
√ √ √ √
ϕ( 2) = ± 2. But ± 2 ∈ / Q( 3). Contradiction. Not isomorphic as fields.
3.8. Order of (a, b) ∈ Z8 × Z4 is lcm(ord(a), ord(b)) = 4. Need pairs with lcm = 4:
elements of order dividing 4 but not all of order dividing 2. Count: elements of order
dividing 4 in Z8 : {0, 2, 4, 6} (4 elements). Elements of order dividing 2: {0, 4}. So
elements of order exactly 4 in Z8 : {2, 6}. In Z4 : ord dividing 4: all 4 elements; ord
dividing 2: {0, 2}; ord exactly 4: {1, 3}. lcm= 4 iff (at least one component has
order 4, or both have order 2 but lcm still 4... lcm(2, 2) = 2 ̸= 4). So: (ord 4 in Z8 ,
any in Z4 ): 2 × 4 = 8. Plus (ord in {1, 2} in Z8 , ord 4 in Z4 ): elements of ord≤ 2 in
Z8 : {0, 4} (2 elements, orders 1 and 2). Pairs: 2 × 2 = 4. Total: 8 + 4 = 12.
3.9. (By cases mod p) x4 + 1 divides x8 − 1. In Zp , by Fermat: ap−1 = 1 for a ̸= 0.
Consider: if p = 2: x4 + 1 = (x + 1)4 in Z2 . Reducible. If p ≡ 1 (mod 8): primitive
8th root of unity exists in Fp , so x4 + 1 has a root. Reducible. If p ≡ 3, 5, 7 (mod 8):
x4 + 1 factors as a product of two quadratics (shown via x8 − 1 = (x4 + 1)(x4 − 1),

15
CSIR-NET Mathematics Appendix C: 350 Practice Problems

and properties of Fp2 ). So x4 + 1 is always reducible over Zp . □

3.10. |G| = 15 = 3 · 5. n5 ≡ 1 (mod 5), n5 | 3: n5 = 1. n3 ≡ 1 (mod 3),
n3 | 5: n3 = 1 or 5. If n3 = 5: 5 × 2 = 10 elements of order 3, plus 4 of order 5,
plus 1 identity = 15. But also need n5 = 1: unique Sylow 5-subgroup P5 (order 5,
|P5 | + |P3 | − |P3 ∩ P5 | = 5 + 3 − 1 = 7 < 15). Contradiction. Actually if n3 = 5:
we’d have 5 × 2 + 4 + 1 = 15, consistent. But standard argument: n5 = 1 so P5 ⊴ G.
Similarly show n3 = 1 by similar counting argument. So both Sylow subgroups are
normal. G = P3 × P5 ∼ = Z3 × Z5 ∼= Z15 . □

16
Units 4–12: Selected Key Problems

Unit 4: Complex Analysis

Problem 4.1 [C∗ ]

Z ∞
x sin x
Evaluate I = 2 2
dx for a > 0.
−∞ x + a

Solution: 4.1
zeiz
I
Consider 2 2
dz over upper semicircle. Poles at z = ±ai; only z = ai is in
C z +a
upper half-plane.
 iz  −a −a
zeiz
Res zze
2 +a2 , ai = limz→ai z+ai = ai·e
2ai
= e2 .
R∞ ix e−a
By Jordan’s Lemma, arc integral → 0. So −∞ xxe 2 +a2 dx = 2πi · 2 = πie−a .
R∞
Taking imaginary part: I = −∞ xx2sin +a2
x
dx = Im(πie−a ) = πe−a · 1 = πe−a .

Problem 4.2 [C∗ ]

If f is entire with |f (z)| ≤ |z|2 + 1 for all z, what can you conclude about f ?

Solution: 4.2
n!·(R2 +1)
By Cauchy’s inequality on |z| = R: |f (n) (0)| ≤ Rn
.
R2 +1
For n ≥ 3: as R → ∞, Rn
→ 0. So f (n) (0) = 0 for all n ≥ 3.
6(R2 +1)
For n = 3: |f ′′′ (0)| ≤ R3
= 6
R
+ 6
R3
→ 0. So f ′′′ = 0 everywhere.
Since all derivatives of order ≥ 3 vanish at 0, f is a polynomial of degree ≤ 2:
f (z) = az 2 + bz + c for some a, b, c ∈ C.

Unit 5: Topology

17
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Problem 5.1 [C∗ ]

Prove that the product of two connected spaces is connected.

Solution: 5.1
Let X and Y be connected. Suppose X × Y = U ∪ V (disjoint open sets).
Fix x0 ∈ X. The slice {x0 } × Y ∼ = Y is connected. Since it intersects U or V (or
both), and it’s connected, it lies entirely in U or entirely in V . WLOG {x0 }×Y ⊆ U .
For any y0 ∈ Y : the slice X × {y0 } ∼
= X is connected and meets U (at (x0 , y0 )). So
X × {y0 } ⊆ U or ⊆ V . Since it contains (x0 , y0 ) ∈ U : X × {y0 } ⊆ U .
So every point of X × Y lies in U , meaning V = ∅. Therefore X × Y is connected.
□

Units 6–12: Rapid Fire Problems

Problems 6.1–12.5 [B] [C∗ ] (Selected Key Problems)

6.1. Solve y ′′ + 4y = cos 2t completely (resonance case). [C∗ ]
6.2. For y ′ = y 2 , y(0) = 1: find the solution and its blow-up time. [B]
7.1. Solve ut = uxx on (0, π) with u(0, t) = u(π, t) = 0, u(x, 0) = x. [C∗ ]
7.2. Classify uxx + 6uxy + 9uyy = 0 and find its characteristics. [B]
8.1. Show ℓ1 ⊊ ℓ2 ⊊ ℓ∞ (proper containments). Give examples in each difference.
[C]
8.2. Prove: If T : H → H is bounded with ⟨T x, x⟩ = 0 for all x, then T = 0. [C∗ ]
√
9.1. Use Newton-Raphson to find 3 7 starting from x0 = 2. Do 2 iterations. [B]
9.2. Find the number of iterations of bisection on [1, 2] needed for error < 10−6 . [B]
R1
10.1. Find the extremal of J[y] = 0 (xy ′ + y 2 ) dx, y(0) = 0, y(1) = 1. [C]
Rx
10.2. Solve the Volterra equation: ϕ(x) = x + 0 (x − t)ϕ(t) dt. [C∗ ]
11.1. For a particle on a sphere of radius R, set up the Lagrangian and find equations
of motion. [C]
11.2. Show {Ly , Lz } = Lx using the Poisson bracket (angular momentum algebra).
[C∗ ]
12.1. Write the dual of: Min 2x1 + 3x2 , s.t. x1 + x2 ≥ 4, x1 − x2 ≤ 2, xi ≥ 0. [B]
12.2. Solve by Simplex: Max z = 3x + 2y, x + y ≤ 4, x − y ≤ 2, x, y ≥ 0. [C∗ ]
12.3. Using complementary slackness, find the dual optimal for 12.2. [C∗ ]

18
CSIR-NET Mathematics Appendix C: 350 Practice Problems

Solution: 6.1–12.3 Solutions

6.1 (Resonance): Char. eq.: r2 + 4 = 0, r = ±2i. yh = A cos 2t + B sin 2t. Since
cos 2t resonates, try yp = t(C cos 2t + D sin 2t). Substituting: −4Ct cos 2t + ...:
4D cos 2t − 4C sin 2t = cos 2t. So D = 1/4, C = 0. yp = t sin4 2t . General:
y = A cos 2t + B sin 2t + t sin4 2t .
1
6.2: Separable: dy/y 2 = dt. −1/y = t + C. y(0) = 1: C = −1. y = 1−t . Blows up
at t = 1.
Rπ n+1 2(−1)n+1 −n2 t
7.1: Bn = π2 0 x sin(nx) dx = 2(−1) . Solution: u = ∞
P
n n=1 n
e sin(nx).
7.2: A = 1, B = 3, C = 9. ∆ = 9 − 9 = 0. Parabolic. Characteristic:
(dy/dx)2 − 6(dy/dx) + 9 = 0 ⇒ dy/dx = 3. Lines y − 3x = C.
P 1 P −3/4 P 2 P −3/2
8.1: xn = 1/n3/4 : xn = n diverges (not in ℓ1 ), xn = n converges
(in ℓ2 ). Shows ℓ1 ⊊ ℓ2 . xn = 1/n1/4 : not in ℓ2 but in ℓ∞ . Shows ℓ2 ⊊ ℓ∞ . □
8.2: From ⟨T x, x⟩ = 0: polarize. ⟨T x, y⟩ = 41 [⟨T (x + y), x + y⟩ − ⟨T (x − y), x − y⟩ +
i⟨T (x + iy), x + iy⟩ − i⟨T (x − iy), x − iy⟩] = 0 for all x, y. Taking y = T x: ∥T x∥2 = 0,
so T x = 0 for all x. T = 0. □
9.1: f (x) = x3 − 7, f ′ (x) = 3x2 . x1 = 2 − 8−7
12
= 2− 1
12
= 23
12
≈ 1.9167.
x2 = 23/12 − (23/12)3 − 7)/(3(23/12)2 ) ≈ 1.9129.
ln(1)−ln(10−6 ) 6 ln 10 13.816
9.2: n ≥ ln 2
= ln 2
≈ 0.693
≈ 20 iterations.
10.1: F = xy ′ + y 2 . Fy = 2y, Fy′ = x. E-L: 2y − dx
d
(x) = 2y − 1 = 0 ⇒ y = 1/2. But
BCs: y(0) = 0 ̸= 1/2. The E-L equation y = 1/2 doesn’t satisfy BCs – this means
the minimum occurs on the boundary or the problem needs natural BCs. Actually
the E-L gives the condition 2y − 1 = 0, a constant – which is a contradiction unless
we’re in a degenerate case. Since Fy′ y′ = 0: the Lagrangian is linear in y ′ , so E-L is
d d
Fy = 0, i.e., 2y = 0... wait: E-L is Fy − dx Fy′ = 2y − dx (x) = 2y − 1 = 0, giving
y = 1/2. But y(0) = 0 fails, so no classical extremal exists within the given BCs.
This is a degenerate variational problem.
Rx
10.2: Differentiate: ϕ′ (x) = 1 + 0 ϕ(t) dt. ϕ′′ = ϕ. ϕ(0) = 0, ϕ′ (0) = 1. ODE
ϕ′′ = ϕ: ϕ = A sinh x + B cosh x. ϕ(0) = B = 0, ϕ′ (0) = A = 1. Solution:
ϕ(x) = sinh x.
11.1: Constrained to sphere: x = R sin θ cos ϕ, y = R sin θ sin ϕ, z = R cos θ.
2
T = mR 2
(θ̇2 + sin2 θ ϕ̇2 ), V = mgR cos θ. L = T − V . E-L for θ: mR2 θ̈ =
mR2 sin θ cos θ ϕ̇2 − mgR sin θ. For ϕ (cyclic): mR2 sin2 θ ϕ̇ = L (const).
11.2: Ly = zpx −xpz , Lz = xpy −ypx . {Ly , Lz } by direct computation using linearity
and {qi , pj } = δij : {zpx − xpz , xpy − ypx } = z{px , xpy } − x{pz , xpy } − z{px , ypx } +
x{pz , ypx } = zpy {px , x} + 0 − 0 + 0 = −zpy (−1) = zpy ... continuing full calculation
gives {Ly , Lz } = ypz − zpy = Lx . □

19
CSIR-NET Mathematics Appendix C: 350 Practice Problems

12.1 (Dual): Max 4y1 + 2y2 , s.t. y1 + y2 ≤ 2, y1 − y2 ≤ 3, y1 ≥ 0, y2 ≤ 0 (from ≥

constraint flipped).
12.2 (Simplex): Optimal at intersection of x + y = 4 and x − y = 2: x = 3, y = 1,
z = 11. (Verified by checking all vertices: (0, 0) → 0, (2, 0) → 6, (4, 0) → 12... wait
x − y ≤ 2 and x ≤ 4: vertex (3, 1): z = 9 + 2 = 11. Vertex (4, 0): z = 12. But check
(4, 0): 4 − 0 = 4 > 2, violates x − y ≤ 2. Infeasible. So optimal is (3, 1), z = 11.
12.3: From optimal tableau, shadow prices y1∗ and y2∗ : dual optimal w = 4y1∗ + 2y2∗ =
11 = z ∗ by strong duality.

20
 CSIR-NET Mathematics Appendix C: 350 Practice Problems

Final 50 Quick-Fire Problems (All Units)

Quick-Fire Problems: Test Yourself

These 50 problems are designed for timed practice (1–2 minutes each for Part B,
5–8 minutes for Part C).

Q1. [B] If {an } is bounded and an+1 ≥ an − 1/n2 , does {an } converge? (Ans: Yes)

Q2. [B] ∞ 1 1 1
P
n=1 n2 +n =? (Ans: 1, telescope n − n+1 )

Q3. [C∗ ] For what values of p does

P p −n
n e converge? (Ans: All p ∈ R)

Q4. [B] The closure of (0, 1) ∪ (1, 2) in R is? (Ans: [0, 2])
R1
Q5. [C∗ ] Is {f ∈ C[0, 1] : 0 f = 0} a closed subspace of (C[0, 1], ∥·∥∞ )? (Ans: Yes)

Q6. [B] If T : Rn → Rm is linear with n > m, must T have a nonzero kernel? (Ans:
Yes, rank-nullity)
 
0 1 0
∗
Q7. [C ] Find the eigenvalues of A = 0 0 1. (Ans: 1, ω, ω 2 , cube roots of unity)
 

1 0 0

Q8. [B] tr(AB) = tr(BA) for all square A, B? (Ans: Yes)

Q9. [C∗ ] Does every group of order 12 have a normal Sylow 3-subgroup? (Ans: Yes,
n3 ∈ {1, 4} but 4 ∤ 4 = 12/3... n3 | 4, n3 ≡ 1 (mod 3): n3 = 1 or 4. Both possible.)

Q10. [B] The number of units in Z/12Z is? (Ans: ϕ(12) = 4)

Q11. [C∗ ] If f is entire and Re(f ) > 0 for all z, must f be constant? (Ans: Yes,
Liouville)

Q12. [B] Res z21+1 , i =? (Ans: 1/(2i))

Q13. [C∗ ] Is [0, 1]ω (countable product of [0, 1]) compact? (Ans: Yes, Tychonoff )

Q14. [B] The Euler-Lagrange equation for J = y ′4 dx is? (Ans: dx d

(4y ′3 ) = 0 ⇒ y ′′ =
R

0)

Q15. [C∗ ] If {Tn } ⊆ B(X) with ∥Tn ∥ ≤ 1 and Tn x → T x for all x: is ∥T ∥ ≤ 1? (Ans:
Yes, lower semicontinuity of norm)

Q16. [B] Order of (2, 3) in Z6 × Z9 ? (Ans: lcm(ord(2), ord(3)) = lcm(3, 3) = 3)

√
Q17. [C∗ ] Solve utt = uxx + u, u(x, 0) = 0, ut (x, 0) = sin x. (Ans: u = sin√ 2t
2
sin x)

21
 CSIR-NET Mathematics Appendix C: 350 Practice Problems

R∞ 2
Q18. [B] 0
xe−x dx =? (Ans: 1/2)

Q19. [C∗ ] If A is n × n with A2 = 0, what are possible values of rank(A)? (Ans:

0, 1, . . . , ⌊n/2⌋)

Q20. [B] Is every subspace of a separable metric space separable? (Ans: Yes)

Q21. [C∗ ] Does Newton-Raphson applied to f (x) = x1/3 starting at x0 ̸= 0 converge?

(Ans: No, diverges since f ′ (0) = ∞)

Q22. [B] ϕ(36) =? (Ans: 12)

Q23. [C∗ ] Is GLn (R) path-connected? (Ans: No, two components: det > 0 and det < 0)

Q24. [B] The Fourier series of f (x) = x2 on [−π, π] gives 1/n2 =? at x = π. (Ans:
P

π 2 /6 via Parseval)
!
0 1
Q25. [C∗ ] Prove the set of nilpotent matrices (over R) is not a subspace. (Ans: +
0 0
!
0 0
is not nilpotent)
1 0

Q26. [B] In Simplex, if all reduced costs are ≤ 0 for maximization, the solution is? (Ans:
Optimal)

Q27. [C∗ ] Characterize all solutions of y ′′ − 4y ′ + 4y = 0. Identify when the trivial

solution is stable. (Ans: y = (c1 + c2 x)e2x . Unstable since eigenvalue r = 2 > 0)
2
Q28. [B] |z|=2 z2z−1 dz =? (Ans: 0; residues at ±1 cancel)
R

Q29. [C∗ ] Prove: ℓ1 is not reflexive (i.e., (ℓ1 )∗∗ ∼

̸= ℓ1 ). (Hint: (ℓ1 )∗ = ℓ∞ , (ℓ∞ )∗ ⊋ ℓ1 )
P n!
Q30. [B] The radius of convergence of z is? (Ans: R = 1)
√ √ √ √ √ √ √ √
Q31. [C∗ ] Is Q( 2 + 3) = Q( 2, 3)? (Ans: Yes, since 2, 3 ∈ Q( 2 + 3))

Q32. [B] Order of S4 ? Number of elements of order 4? (Ans: 24; elements of order 4
are 4-cycles: 4!4 = 6)

Q33. [C∗ ] For = p2 /2m − k/r (Kepler): show {L2 , } = 0 (L2 = L2x + L2y + L2z ). (Angular
momentum conserved in central force)

Q34. [B] Find an LPP that is feasible but unbounded. (Ans: Max x, s.t. x ≥ 0)

Q35. [C∗ ] Prove the open mapping theorem implies the closed graph theorem.
(Ans: If graph Γ(T ) is closed in X × Y , consider S : Γ(T ) → X, (x, T x) 7→ x,
which is bijective bounded. By OMT, S −1 is bounded, so T is bounded.)

22
 CSIR-NET Mathematics Appendix C: 350 Practice Problems

 
1 2 3
Q36. [B] det 4 5 6 =? (Ans: 0, rows are linearly dependent)
 

7 8 9

Q37. [C∗ ] Prove or disprove: every continuous map f : S 1 → S 1 has a fixed point. (Ans:
False. Rotation by π has no fixed point.)

Q38. [B] Simpson’s rule is exact for polynomials of degree ≤? (Ans: ≤ 3)

Q39. [C∗ ] State and give one application of Burnside’s Lemma. (Ans: Number of distinct
necklaces with n beads of k colors = n1 d|n ϕ(n/d)k d )
P

Q40. [B] The kernel of a ring homomorphism is always a? (Ans: Two-sided ideal)
R1
Q41. [C∗ ] For T : L2 [0, 1] → L2 [0, 1], T f = 0 f dx (constant function). Find ∥T ∥,
eigenvalues, and spectrum. (Ans: ∥T ∥ = 1. Eigenvalue λ = 1 (eigenfunctions:
R
constants). λ = 0 (eigenspace: {f : f = 0}). Spectrum = {0, 1})
ln n
Q42. [B] limn→∞ n
=? (Ans: 0, L’Hôpital)

Q43. [C∗ ] Prove: every bounded sequence in a Hilbert space has a weakly convergent
subsequence. (Ans: Use sequential compactness of unit ball under weak topology,
or Banach-Alaoglu)

Q44. [B] The dual of the dual LPP is the? (Ans: Original primal LPP)

Q45. [C∗ ] Show xn = 11 + 12 + · · · + n1 − ln n converges (Euler-Mascheroni constant). (Ans:

Show monotone decreasing and bounded below by 0)

Q46. [B] If f is analytic on |z| < 2 and f (n/(n + 1)) = 0 for all n, then f ≡ 0? (Ans:
Yes, identity theorem: zeros accumulate at 1)

Q47. [C∗ ] Prove: every complete metric space is a Baire space (countable intersection
of open dense sets is dense). (Ans: Standard Baire Category Theorem proof )

Q48. [B] The general solution of y ′ = 2y/x is? (Ans: y = Cx2 )

Q49. [C∗ ] Let V, W be finite-dim over R. Prove dim(V ⊗ W ) = dim V · dim W . (Ans:
If {ei }, {fj } are bases, {ei ⊗ fj } is a basis for V ⊗ W )

Q50. [B] The Wronskian of ex and e−x is? (Ans: W = ex (−e−x ) − e−x (ex ) = −2)

23
CSIR-NET Mathematics Appendix C: 350 Practice Problems

You have completed Appendix C!

350+ problems across all 12 units have been presented with
complete solutions.

Revisit problems you struggled with. Rework them without hints.

Consistent practice is the bridge between knowledge and JRF.

Best wishes for your CSIR-NET examination!

24