Analysis I

Prof. Tamás Hausel

Fall Semester 2015–2016 ÉC O L E P O L Y T E C H N I Q U E
Posted December 18, 2015 FÉ DÉR A L E D E L A U S A N N E

Course Review Solutions

n 5
1. (a) lim = lim 1 − = 1.
n→+∞ n + 5 n→+∞ n+5
n2 + 2 1 2 1
(b) lim = lim + = .
n→+∞ 3n2 + 4 n→+∞ 3 3(3n2 + 4) 3
√ p
n2 + 5 n 1 + 5/n2 1
(c) lim = lim = .
n→+∞ 3n n→+∞ 3n 3
√
cos n 3 √
(d) lim = 0 because cos n3 is bounded.
n→+∞ n
√ 1
(e) lim n sin = 0 because 0 ≤ sin n1 ≤ n1 hence the conclusion by the squeeze theorem.
n→+∞ n
1 1
(f) lim n cos sin 2 = 0 noticing that cos is bounded and the inequalities for sin above.
n→+∞ n n
sin(n + 2) − sin(n − 2) 2 cos n sin 2
(g) lim = lim = tan 2.
n→+∞ cos(n + 2) + cos(n − 2) n→+∞ 2 cos n cos 2

cos(n + 1) + cos(n − 1) 2 cos n cos 1

(h) lim = lim = 2 cos 1.
n→+∞ cos n n→+∞ cos n
√
sin n2 + 5
(i) lim 3 = 0 since sin is bounded.
n→+∞ n + n2 + 1
√ √
n2 + 2 − n2 + 3 −1
(j) lim = lim √ √ = 0.
n→+∞ 5 n→+∞ 5( n2 + 2 + n2 + 3)
p p  −9n − 1 9
(k) lim 2n2 + 3 − (2n + 1)(n + 4) = lim √ p = − √ =
n→+∞
√
n→+∞ 2n2 + 3 + (2n + 1)(n + 4) 2 2
9 2
− .
4
p n(6n + 3)
(l) lim n( n4 + 6n + 3 − n2 ) = lim √ = 3.
n→+∞ n→+∞ n4 + 6n + 3 + n2
√ p 3 p
(m) lim n( n + 2n − n3 + 4) = 1 as in (k) above.
n→+∞

n3
(n) lim cos n2 = 0 as cos is bounded and n < A·6n/3 for all n ∈ N and A = (61/3 −1)−1 .
n→+∞ 7n
2n
(o) lim = 0 since, for example, n! > 3n for all n > 3 (proof by induction).
n→+∞ n!
2
(p) lim 2n e−2n = 0 noticing that 2 < 1.
n→+∞ e

n
n  n+1
2
(q) lim 1 + n2 = lim 1 + n1 1 + n+11 n+1
n+2 = e .
n→+∞ n→+∞
 n
1 n−1 1 1
(r) lim 1− = lim 1 = .
n→+∞ n n→+∞ n (1 + n−1 )n−1 e
n n n
(s) lim 1 − n12 = lim 1 − n1 1 + n1 = 1.




n→+∞ n→+∞

1
 (t) lim n3 (1 − cos n1 ) sin n1 = 12 . We use the fact that
n→+∞

sin n1 sin x
lim n sin n1 = lim 1 = lim =1 and lim cos n1 = 1
n→+∞ n→+∞ x→0 x n→+∞
n

since
1 1 1
1 > n sin > cos > 1 − .
n n n
To transform 1 − cos n1 there are two possibilities. First, writing

1 − cos2 n1 sin2 n1
1 − cos n1 = =
1 + cos n1 1 + cos n1
or by using the identity

1 − cos n1 = cos 0 − cos n1 = 2 sin2 1

2n .

With the first result, we have

sin2 n1 n3 sin3 n1 1
lim n3 (1 − cos n1 ) sin n1 = lim n3 1 sin 1
n = lim 1 = 2.
n→+∞ n→+∞ 1 + cos n n→+∞ 1 + cos n
If we use the second identity, we have

lim n3 (1 − cos n1 ) sin n1 = lim n3 2 sin2 ( 2n

1
) sin n1 = 1
lim (2n)3
2 n→+∞ sin3 ( 2n
1 1
) cos 2n = 12 .
n→+∞ n→+∞

1 1
2. (a) For x 6= 0, lim 2 2 = 0. If x = 0, then lim 2 2 =1
n→+∞ 1+n x n→+∞ 1+n x
2 3 2 3
1−n x
(b) For x 6= 0, lim = −x. If x = 0, then lim 1−n
2 2
x
2 2 = 1
n→+∞ 1+n x n→+∞ 1+n x
 n 2
x −1
(c) For x 6= −1, 1, lim n = 1. If x = 1 the limit is 0.
n→+∞ x +1
√ √
(d) For all x > 0, lim n( n x − 1) = 0 (see ch. 2.3 from lectures).
n→+∞

3. (a) Note that

n n (n+1)!
Y k2 − 1 Y (k + 1)(k − 1) 6 (n − 1)! 4(n − 1)
xn = = = =
k2 − 4 (k + 2)(k − 2) (n+2)!
− 2)! n+1
k=3 k=3 24 (n

hence
∞
Y k2 − 1
lim xn = = 4.
n→+∞ k2 − 4
k=3

(b) We proceed by induction. If n = 2, then

23 − 1 2 22 + 2 + 1
x3 = 3
= · .
2 +1 3 2(2 + 1)
For the induction step, we use the factorization a3 − b3 = (a − b)(a2 + ab + b2 )) and
observe
(n + 1)3 − 1 2 n2 + n + 1 (n + 1)3 − 1
xn+1 = xn · = · ·
(n + 1)3 + 1 3 n(n + 1) (n + 1)3 + 1
2 n2 + n + 1 n((n + 1)2 + (n + 1) + 1) 2 (n + 1)2 + (n + 1) + 1
= · · = ·
3 n(n + 1) (n + 2)((n + 1)2 − (n + 1) + 1) 3 (n + 1)(n + 2)

2
 Finally, we use the above to compute the limit
∞
Y k3 − 1 2
lim xn = = .
n→+∞ k3 + 1 3
k=2

n2
4. (a) lim = ∞.
n→+∞ n + 2 + (−1)n
n
(b) lim = ∞.
n→+∞ cos 1
n
n! n! 2 3n
(c) lim n = ∞ since for all n ≥ 3 we have n! ≥ 2 · 3n−2 hence 2n ≥ 9 · 2n .
n→+∞ 2

(2+3i)n+5 2+3i
5. (a) lim = =i
n→∞ (3−2i)n+1 3−2i

(b) lim (7+10i)n+1 = 148+105i

n→∞ (14+5i)n+2 221
(3+9i)n+3 3
(c) lim =
n→∞ (4+12i)n+7 4
√ r√
√
r
√ 3+ 32 +42 32 +42 −3
(3+4i)n2 +(5+7i)n−2 3+4i +i
2 2 2+i
(d) lim (2+i)n+3 = 2+i = 2+i = 2+1 =1
n→∞
√ r√
√
r
√ 5+ 52 +122 52 +122 −5
(5+12i)n2 +(11−3i)n+8 5+12i +i
2 2 3+2i
(e) lim (2−3i)n−7 = 2−3i = 2−3i = 2−3i =i
n→∞
s √ s√
√ √ 8+ 82 +152
+i
82 +152 −8 √
(8+15i)n2 +(2+7i)n+12 8+15i 2 2 2(31−29i)
(f) lim = = = √ 5+3i =
n→∞ (2+7i)n−9 2+7i 2+7i 2(2+7i) 106

ei(n+a) −e−i(n−a) eia (ein −e−in ) 2ieia sin n

(g) lim sin n = lim sin n = lim sin n = 2ieia
n→∞ n→∞ n→∞
n(e2i(n+a) +e2i(a−n) ) ne2ia (e2in +e−2in ) 2n 2ia
(h) lim (2in+1) cos 2n = lim = e2ia lim = e i = −ie2ia
n→∞ n→∞ (2in+i) cos 2n n→∞ 2in+1
1 1 √ iθ
(i) Using the polar expression, lim (a + ib) n = lim (reiθ ) n = lim n re n = 1 · e0 = 1.
n→∞ n→∞ n→∞
(j) Using the polar representation, we may write
p
a + ibn = a2 + b2 n2 eiθn
√ θn
for some θn ∈ [−π, π[. Then we have lim n a2 + b2 n2 = 1 and lim = 0. Hence
n→∞ n→∞ n
1
p
n iθn
lim (a + ibn) n = lim a2 + b2 n2 e n = 1.
n→∞ n→∞
ia
(k) Expanding the exponential, we obtain n(e n − 1) = n(cos na − 1) + in sin na . We recall
the following trigonometric inequality derived in Chapter 1: for all 0 < |h| < 1, we have
sin h
1 − h2 < cos h < < 1.
h
a
We set h = n. For all n > |a|, we have 0 < |h| < 1 and so
a2 a a2
1− 2
< cos < 1 ⇒ − < n(cos na − 1) < 0 ⇒ lim n(cos na − 1) = 0.
n n n n→∞

Similarly,
a2 n a a
1− < sin < 1 ⇒ lim n sin = a.
n2 a n n→∞ n
We then conclude that
ia
lim n(e n − 1) = ia.
n→∞

3
 (l) Using the preceding result,
ia ib
 ia ib

lim n(e n − e− n ) = lim n(e n − 1) − n(e− n − 1) = i(a + b).
n→∞ n→∞

Pn 1 3
6. (a) We have lim k=2 k2 −1 = 4 since
n→+∞

n n   n n
!
X 1 X 1 1 1 1 1 X 1 X 1
1

1 1 1

= − = − = 2 1+ 2 − n − n+1 .
k2 − 1 2k−1 2k+1 2 k−1 k+1
k=2 k=2 k=2 k=2

Pn 1 1
(b) We have lim k=0 k2 +5k+6 = 2 since
n→+∞

n n n
X 1 X 1 X 1 1 1 1
2
= = − = − .
k + 5k + 6 (k + 2)(k + 3) k+2 k+3 2 n+3
k=0 k=0 k=0

Pn k+2
(c) We have lim k=0 k3 +5k2 +8k+2 = 1 since
n→+∞

n n n
X k+2 X 1 X 1 1 1
= = − =1− .
k 3 + 5k 2 + 8k + 2 (k + 1)(k + 2) k+1 k+2 n+2
k=0 k=0 k=0

 
1
Pn Pn n(n+ 12 )(n+1)
(d) We have lim n3 = 31 since k=1 k 2 =
k=1 k2 3 .
n→+∞
 
1
Pn 2
(e) Using the above, we find lim (n+1)! k=1 k = 0.
n→+∞

k2 +4k+3
Pn 3
(f) One has lim k=0 k4 +8k3 +22k2 +24k+9 = 4 since with Exercise (a)
n→+∞

n n n+2
X k 2 + 4k + 3 X 1 X 1
4 3 2
= = .
k + 8k + 22k + 24k + 9 (k + 1)(k + 3) k2 − 1
k=0 k=0 k=2

P+∞
(g) This is a geometric series: n=0 51n = 54 .
P+∞  
(h) One has n=1 sin n1 − sin n+1 1
= sin 1 − sin 0 = sin 1.
P+∞  
(i) We find n=1 √1n − √n+1 1
= 1 since for all convergent sequences (xn ) ,

N
X
lim (xn − xn+1 ) = lim (x1 − xN +1 ) = x1 − lim xN .
N →+∞ N →+∞ N →+∞
n=1

P+∞
(j) The formula for geometric series gives n=0 e−n = e
e−1 .

1 1
7. (a) The series converges since for all n ∈ N, 0 < n2 +n+3 < n2 .
2n2 +1
(b) The series diverges since 3n2 +2 does not converge to 0 as n → ∞.
n
(c) The series diverges since (−1)n n+3 does not converge to 0 as n → ∞.
sin 2n2 1
(d) The series converges since for all n ∈ N, n2 +3 < n2 .

sin n! 1
(e) The series converges since for all n ∈ N, n2 +1 < n2 .
cos 2n 1
(f) The series converges since for all n ∈ N, n2 < n2 .

4
 √ √
n+4− n+2 2 √ 1
(g) The series converges since for all n ∈ N, 0 < n = √
n( n+4+ n+2)
< n3/2
.
1 1
(h) The series converges since for all n ∈ N, 0 < √ < n3/2 .
n(n2 +3)
√ √
5 n5
(i) The series diverges since for all n ∈ N, n3n+1 > n3 +n 1
3 = 2n1/2 .
√
(j) The series converges since for all n ∈ N, 0 < n2 +2n√n < n3/2
1
.
(k) The series converges thanks to Leibniz convergence criteria for alternating series. The
1
sequence √n+3 is decreasing and converges to 0.
(l) The series converges thanks to Leibniz convergence criteria for alternating series. The
sequence √n12 +1 is decreasing and converges to 0.
(m) The series converges thanks to Cauchy’s criterion since
q √
n √
lim n n
n−1 = lim n n − 1 = 0.
n→+∞ n→+∞

8. If α > 1 the series converges since for all integers n > 0:

αn αn α 1
0< < 2n =
α2n −1 α − α2n−1 α − 1 αn
or by Cauchy’s criterion as
r
n αn 1
lim = < 1.
n→+∞ α2n −1 α
1
If 0 < α < 1, the series also converges. We observe that for n large enough, α2n < 2 and so

αn
< 2αn ,
α2n −1
so the series is absolutely convergent. Alternatively, using Cauchy’s criterion and observing
that
s
αn
lim n = α < 1.
n→+∞ α2n − 1

9. (a)
√
9+x−3 1 1
lim = lim √ =
x→0 x x→0 9+x+3 6

(b) Noting that |x| = −x for x < 0 we find

√
q
3
x2
+3 |x| 1 + x2 1
lim = lim =− .
x→−∞ 4x + 1 x→−∞ 4x + 1 4

(c)


p x x2 + 2 − x2


lim x x2 + 2 − x = lim √
=1
x→+∞ x→+∞ x2 + 2 + x

(d)
√ √ q  √ √
3x + 1 − 2x + 1 π π( 3 − 2)
q
1 1
lim √ arctan x = lim 3+ x − 2+ x =
x→+∞ x 2 x→+∞ 2

5
(e)
√ √ √ √ √ √ √ √
2 + x − 2x ( 2 + x − 2x)( 2 + x + 2x)( 1 + 2x + 5)
lim √ √ = lim √ √ √ √ √ √
x→2 1 + 2x − 5 x→2 ( 1 + 2x − 5)( 1 + 2x + 5)( 2 + x + 2x)
√ √ √ √
(2 − x)( 1 + 2x + 5) 1 1 + 2x + 5 5
= lim √ √ = − lim √ √ =−
x→2 2(x − 2)( 2 + x + 2x) 2 x→2 2 + x + 2x 4

(f)
√ √ √
x− 3+ x+3
lim √ = +∞
x→3+ x2 − 9

(g)

x 1 + x2 4 + x3 − 4
q 



2

3

11
lim x 1+ x 4+ x − 2 = lim q =
x→+∞ x→+∞
1 + x2 4 + x3 + 2


4

(h)
√
q p
x+ x+ x
lim √ =1
x→+∞ x+2

(i)

x3 + x2 − 5x + 3 (x − 1)2 (x + 3) x+3 4
lim = lim = lim = .
x→1 x3 − 3x + 2 x→1 (x − 1)2 (x + 2) x→1 x + 2 3

(j)

x4 − x3 + x2 − 1 x3 + x + 1 3
lim 2
= lim =
x→1 x −1 x→1 x+1 2
(k)

3x3 + 2x2 + x + 1
lim =3
x→+∞ x3 − x

(l)
 
1 3 x−2
lim − = lim = −1
x→−1 1 + x 1 + x3 x→−1 x2 −x+1

(m) lim [x] does not exist since [x] = 0 if x > 0 and [x] = −1 if x < 0.
x→0
x x x
(n) The inequalities [x] ≤ x < [x]+1 imply that ≤ < 1 hence lim 1+[x] = 1.
1+x 1 + [x] x→+∞
x
(o) lim [x] does not exist since the function is not defined for 0 < x < 1.
x→0
(p)
sin x − sin(x − π) − sin h
lim = lim = lim = −1
x→π x − π x→π x−π h→0 h
(q)
tan x sin x
lim = lim =1
x→0 x x→0 x cos x

6
 (r) lim sin x =0
x→+∞ x

(s) lim x sin x1 = lim sin h

h =1
x→+∞ h→0+

(t) lim cos x1 does not exist (see lecture notes).

x→0
sin
√x
(u) lim x
=0
x→0+


(v) lim ln(1 + a sinh x) − ln(cosh x) = ln a, a > 0.
x→+∞
2
(w) lim x6 e−x = 0
x→+∞

(x) lim x3 3−x = 0

x→+∞

10. (a) Either using the definition of the derivative

ex − 1 d ex
lim = =1
x→0 x dx x=0

or by l’Hospital’s rule (it is an indefinite expression - see lecture notes; the derivatives
of the numerator and the denominator exist and limit in the resulting expression can
be evaluated)

ex − 1 ex
lim = lim = e0 = 1.
x→0 x x→0 1

In the rest of the solution, we won’t verify explicitly the condition of validity of
l’Hospital’s rule. But in an exam, such a verification is compulsory.
(b)
3 3
ex − 1 3xex
lim = lim =0
x→0 x2 x→0 2

(c)

cos 2x −2 sin 2x
lim = limπ = −2.
x→ π
4 x − π4 x→ 4 1

(d)
p √ √ √
xx − 3 3 ex ln x − e 3 ln 3 (1 + ln x)ex ln x 1
p √
lim
√ √ = lim
√ √ = lim
√ = (1 + 2 ln 3) 3 3
x→ 3 x − 3 x→ 3 x− 3 x→ 3 1

(e) lim sin x cos x

x−π = lim 1 = −1
x→π x→π

(f) lim sinx x = lim cos1 x = 1

x→0 x→0
x+sin x 1+cos x
(g) lim 1−cos x = lim sin x and this limit does not exist.
x→0 x→0

(h) lim x = lim 1

1 =1
x→0 arctan x x→0 1+x2

(i)

sin x sin x cos x

lim
= lim = lim = −1
x→0 1
ln 1+x x→0 − ln(1 + x) x→0 − 1
1+x

7
 1
(j) If α > 0, then writing x = y, Example 7 from Chapter 4.5 gives lim xα ln x =
x→0+
− lim y −α ln y = 0.
y→+∞
ln x
(k) For all α > 0, lim α = 0, by Example 7 from Chapter 4.5.
x→∞ x
(l) By continuity of power functions, lim (1 + αx )α = lim (1 + y)α = 1.
x→∞ y→0

(m) Either by Example 2 from Chapter 4.5, or by n successive uses of l’Hospital’s rule,
x x
lim xen = lim en! = +∞.
x→∞ x→∞
(n) By a (long and) direct calculation using l’Hospital’s rule

arctan 1−x − 12


1+x 1
lim = lim 1+x = − .
x→1 x−1 x→1 1 2
1−x −2y
Alternatively, we write y = 1+x , so x − 1 = 1+y and
1−x

y+1
arctan 1+x −(y + 1) arctan y − arctan y − 1+y 2 1
lim = lim = lim =−
x→1 x−1 y→0 2y y→0 2 2
2
11. (a) One has truncated Taylor expansions cosh x = 1 + x2 + O(x4 ) and ln(1 + y) = y + O(y 2 )
and hence
x2
 2
x4
  
2 2 4 2 x 4
x ln cosh x = x ln 1 + + O(x ) = x ( + O(x ) = + O(x6 )
2 2 2

and ln2 (1 + x2 ) = x4 + O(x5 ). We then find

x2 ln cosh x 1
lim 2 = .
x→0 2
ln (1 + x ) 2

(b) One has truncated Taylor expansions sinh x = x + O(x3 ) and ln(1 + y) = y + O(y 2 ) so

ln(1 + x2 )
lim =1
x→0 sinh2 x
(c) By l’Hospital’s rule
√
1 + x2 x tanh x
lim = lim √ = 1.
x→∞ ln sinh x x→∞ 1 + x2

(d) l’Hospital’s rule gives

sin √2x −2x−3/2 cos √2x

lim √ √ = lim 1√ = −2.
x→∞ ln( x − 1) − ln x x→∞ x− x
− x1

Alternatively, using truncated Taylor expansions with y = √1 , we find

x

sin √2x sin 2y

lim √ √ = lim = −2
x→∞ ln( x − 1) − ln x y→0 + ln(1 − y)
cos( πx
2 ) sin(x−1)
(e) lim ln((x−1)2 ) =0
x→1
(f)

cos x − cos x1 1
lim 1 =0 since e x goes to infinity and the other terms are bounded.
x→0+ ex − e x

8
 (g) Using truncated Taylor series for sin x and cos x,
2
x 4
1 + cos(x − π) 1 − cos x 2 + O(x ) 1
lim = lim = lim = .
x→0+ x sin x x→0+ x sin x x→0+ x(x + O(x3 )) 2
2
12. (a) f (x) = xe−x is of class C ∞ with
2 2
f 0 (x) = (1 − 2x2 )e−x f 00 (x) = −2x(3 − 2x2 )e−x .
√ √ 1
√
Stationary points: x1 = − 22 (strict local minimum, f (x1 ) = − 22 e− 2 ) and x2 = − 2
2
√ 1
(strict local maximum, f (x2 ) = 22 e− 2 ). These are in fact global extrema since

lim f (x) = 0.
x→±∞
√
6
Inflexion points: x = 0 x = ± 2 .
(b) On the boundary f (−1) = f (3) = 0 and

lim f (x) = ∞.
x→±∞

More precisely, one observes that

lim f (x) − (1 − x) = 0 lim f (x) − (x − 1) = 0

n→−∞ n→∞

so that there are asymptotes y = 1 − x when x → −∞ and y = x − 1 when x → +∞.

f is of class C ∞ on the open set ] − ∞, −1[∪]3, ∞[ with
x−1 −4
f 0 (x) = p f 00 (x) = p .
(x + 1)(x − 3) ( (x + 1)(x − 3))3

f is strictly monotonic and strictly concave on its domain.

13. (a) We note that f (x) = exp(x ln x − x) and so f 0 (x) = (ln x)f (x). Hence x = 1 is the
unique stationary point of the function f as f (x) > 0 and ln x is strictly increasing.
(b) f has a strict local minimum at x = 1 since f 0 (x) < 0 if x < 1 and f 0 (x) > 0 if x > 1.
The truncated Taylor expansion of order 4 at this point is given by

f (x) = e−1 1 + 21 (x − 1)2 − 16 (x − 1)3 + 24

5
(x − 1)4 .

(c) Using that x ln x → 0 when x → 0 (see Chapter 4 in the lecture notes), we have, by the
continuity of the exponential function,

lim f (x) = 1.
x→0+

By l’Hospital’s rule (one may verify the required conditions are satisfied) we get

f 0 (x)
  
f (x) − 1 (ln x)f (x) ln x
lim = lim = lim = lim lim f (x)
x→0+ x ln x x→0+ 1 + ln x x→0+ 1 + ln x x→0+ 1 + ln x x→0+
 1  
= lim x1 lim f (x) = 1.
x→0+ x→0+
x

14. (a)
Z x
cos t
dt = ln(1 + sin x)
1 + sin t

9
(b)
√
Z x √ Z x+2 √ √
t+2
e dt = 2s es ds = 2( x + 2 − 1)e x+2

(c) Note that

x
e2t − 1 1 − e−t x + e−x
Z
sinh t 1 sinh t
t
= e−t t = so t
dt =
e +1 2 e +1 2 e +1 2

(d)
Z x Z ex
1 1
dt = 2 ds = 2 arctan(ex ).
cosh t 1 + s2

(e) Note that ex = cosh x + sinh x so

Z x x
et
Z
sinh x
dt = 1+ dt = x + ln(cosh x)
cosh t cosh t

(f)
x x x
ln(1 + x2 )
Z Z Z
t
arctan t dt = (t)0 arctan t dt = x arctan x − 2
dt = x arctan x −
1+t 2

(g)
√ √ √
Z x √ Z x Z x
2 0
√ s2
Z x
arctan( t) dt = 2s arctan s ds = (s ) arctan s ds = x arctan( x) − ds
1 + s2
√
√
Z x √ √ √
1
= x arctan( x) − 1− ds = x arctan( x) − x + arctan( x)
1 + s2

(h)
Z x Z x Z x
2t arccos t
arccos2 t dt = (t)0 arccos2 t dt = x arccos2 x + √ dt
1 − t2
Z x p
= x arccos2 x + 2 (− 1 − t2 )0 arccos t dt
Z x√
p 1 − t2
= x arccos2 x − 2 1 − x2 arccos x − 2 √ dt
1 − t2
p
= x arccos2 x − 2 1 − x2 arccos x − 2x

(i) Note that 1 + tan2 t = cos−2 x, so

sin t sin t cos2 t

=
1 + cos t + tan2 t 1 + cos3 t
and hence
Z x
sin t 1
dt = − ln(1 + cos3 x)
1 + cos t + tan2 t 3

(j) By integration by parts and Exercise (l) below

Z x Z x x
x2 cos x2 −x2 cos x2 + sin x2
Z
t3 sin t2 dt = − t2 ( 12 cos t2 )0 dt = − + t cos t2 dt = ;
2 2

10
 or by a change of variable and integration by parts
x x2 Z x2 x2
s(− cos s)0 x2 cos x2
Z Z Z
3 2 s sin s cos s
t sin t dt = ds = ds = − + ds
2 2 2 2
−x2 cos x2 + sin x2
= .
2

(k) Integrating by parts twice successively ,

Z x Z x Z x Z x
I := t2 cos2 t dt = t2 cos t(sin t)0 dt = x2 cos x sin x − 2 t cos t sin t dt + t2 sin2 t dt
Z x Z x
2 2 0
= x cos x sin x − t(sin t) dt + t2 (1 − cos2 t) dt
Z x Z x
= x2 cos x sin x − x sin2 x + sin2 t dt + t2 dt − I

x − sin x cos x x3
= x2 cos x sin x − x sin2 x + + −I
2 3
Rx 2
with the result of the lectures for sin t dt. Hence,
Z x
x − sin x cos x x3
 
1
t2 cos2 t dt = x2 cos x sin x − x sin2 x + + .
2 2 3

(l) By the change of variable s = t2 ,

x x2
sin x2
Z Z
cos s
t cos t2 dt = ds = .
2 2

(m) Integrands containing the trigonometric functions R(sin t, cos t, tan t, cot t) can be trans-
formed into algebraic integrands by the change of variable

z = tan 2t .

With this, we have

2z 1 − z2
sin t = 2 sin 2t cos 2t = 2 tan 2t cos2 t
2 = cos t = cos2 t
2 − sin2 t
2 =
1 + z2 1 + z2
2z dt 2
tan t = = .
1 − z2 dz 1 + z2
So
Z x x
tan
1 − z2 1 − z2
Z  
2 2z 2z 2
R(sin t, cos t, tan t, cot t) dt = R , , , dz
1 + z2 1 + z2 1 − z2 2z 1 + z2

In particular,
x x
x tan tan
1 + z2
Z Z Z
1 2 2 2 1 x
dt = · dz = dz = − ln tan − 1 .
cos t − sin t + 1 2(1 − z) 1 + z 2 1−z 2

15. (a) Integrating by parts twice (see lecture notes), we find that
Z x
ex (sin x + cos x)
et cos t dt = .
2

11
(b) With the change of variable s = ln t (equivalently, t = es ) and the previous exercise
Z x Z ln x
x(sin(ln x) + cos(ln x))
cos(ln t) dt = es cos s ds = .
2

(c)
x x x x
x2 ln x 1 2x2 ln x − x2
Z Z Z Z
2 1 2 0
t ln t 2/3
dt = t ln t dt = (t ) ln t dt = − t2 · t−1 dt =
3 3 3 3 6

(d)
x x x
(1 + t)3
Z Z Z
1
(1 + t)2 ln t3 dt = ((1 + t)3 )0 3 ln t dt = (1 + x)3 ln x − dt
3 t
x3 3x2
= (1 + x)3 ln x − ln x − − − 3x
3 2

(e)
x
1 x 3 0 2 x3 ln2 x 2 x 2 x3 ln2 x 2 x 3 0
Z Z Z Z
t2 ln2 t dt = (t ) ln t dt = − t ln t dt = − (t ) ln t dt
3 3 3 3 9
x3 ln2 x 2x3 ln x 2 x 2 x3 ln2 x 2x3 ln x 2x3
Z
= − + t dt = − +
3 9 9 3 9 27

(f)
Z x Z x Z x
ln3 t dt = (t)0 ln3 t dt = x ln3 x − 3 ln2 t dt = · · · = x ln3 x − 3x ln2 x + 6x ln x − 6x

(g)
x
1 x 2 0 x2 ln(2 + x) 1 x t2
Z Z Z
t ln(2 + t) dt = (t ) ln(2 + t) dt = − dt
2 2 2 2+t
x2 ln(2 + x) 1 x x2 ln(2 + x) − 4 ln(2 + x) x2
Z
4
= − t−2+ dt = − +x
2 2 2+t 2 4

(h)
Z x Z x Z x
ln t ln x 1
dt = − ((1 + t)−1 )0 ln t dt = − + dt
(1 + t)2 1+x t(1 + t)
Z x
ln x 1 1 x ln x
=− + − dt = − ln(1 + x)
1+x t 1+t 1+x

(i) Either by the change of variable of Exercise 13 given by z = tan 2t , so that sin t = 1+z
2z
2,

in which case we obtain

Z x Z tan x2 Z tan x2
sin t 4z 2 2
dt = dz = − dz
1 + sin t (1 + z)2 (1 + z 2 ) 1 + z2 (1 + z)2
x
tan
2 2
2
= + 2 arctan z = +x
1+z 1 + tan x2

12
 or by integrating by parts
Z x Z x Z x
sin t (− cos t)0 − cos x cos2 t
dt = dt = − dt
1 + sin t 1 + sin t 1 + sin x (1 + sin t)2
Z x Z x x
− cos x 1 − sin t − cos x
Z
1 sin t
= − dt = − dt + dt
1 + sin x 1 + sin t 1 + sin x 1 + sin t 1 + sin t
whence we deduce that
x
− cos x
Z
1
dt =
1 + sin t 1 + sin x
and hence that
Z x Z x
sin t 1 cos x
dt = 1− dt = x + .
1 + sin t 1 + sin t 1 + sin x

(j)
x
cos9 x
Z
sin t cos8 t dt = −
9

(k)
Z x Z x Z x
sin 2t 2 sin t cos t 4 sin t
dt = dt = 2 sin t − dt = −2 cos x + 4 ln(2 + cos x)
2 + cos t 2 + cos t 2 + cos t

(l) Using the change of variable s = cos t, we find

Z x Z cos x
sin t 1
dt = − ds
2 + cos2 t 2 + s2
√
A second change of variable s = 2z (ideally, we would do the two changes of variable
simultaneously) gives us
Z x Z cos
√ x
√ √ √ !
sin t 2 2 2 2
dt = − dz = − arctan cos x
2 + cos2 t 2(1 + z 2 ) 2 2

(m) By the change of variable s = cos t,

Z x Z cos x
cos2 t sin t s2
dt = − ds = − cos x + arctan(cos x).
1 + cos2 t 1 + s2

(n) By the change of variable t = s2 ,

Z x √ √
x √
Z
t 1
dt = 2 2
ds = 2 arctan x
t(1 + t) 1+s

(o)
Z x √ 2 3 2 9
(1 + t3 ) t dt = x 2 + x 2
3 9

(p) By the change of variable t3 = s,

x x3 Z 3
√ 1 x
Z Z
5
p 1
t 1− t3 dt = s 1 − s ds = (1 − s)1/2 − (1 − s)3/2 ds
3 3
2(1 − x3 )3/2 2(1 − x3 )5/2
=− +
9 15

13
 (q) By the change of variable t = s2 − 1,
x √ x
√
x+1
√
Z Z Z
t+1 1 1 2
dt = √ + √ dt = 2 x + 1 − ds
t t+1 t t+1 1 − s2
√ √
= 2 x + 1 − 2 arctanh x + 1

(r) By integration by parts,

Z x√ Z x √ x
√
1 + t2 − 1 + x2 − 1 + x2
Z
−1 0
p 1
dt = (−t ) 1 + t 2 dt = + √ dt = + arcsinh x
t2 x 1 + t2 x

(s)
x x
x − 3

Z Z
1 1
p dt = p dt = arcsin
t(6 − t) 9 − (t − 3) 2 3

(t)
Z x
1 3 p
√ dt = ln x − + x2 − 3x
t2 − 3t 2

16. (a)
Z 1 √ 1 1 2 0
Z
1 1
Z
1
t ln t dt = (t ) ln t dt = − t dt = −
0 4 0 4 0 8

(b)
Z +∞
1
dt = lim arctan x − lim arctan x = π
−∞ 1 + t2 x→+∞ x→−∞

(c)
Z +∞ Z +∞ Z +∞
1 1 1
dt = dt = dt = π
−∞ t2 + 8t + 17 −∞ (t + 4)2 + 1 −∞ 1 + t2

(d)
+∞ +∞
3t − 1
Z Z
3 1 4t
dt = − + 2 dt
1
2
t(4t2 + 1) 1
2
4t2 +1 t 4t + 1
∞ ∞
3 1
= arctan(2t) + (− ln t + ln(4t2 + 1))
2 1 2 1
2 2
∞ 2 ∞
3 1 4t + 1
3π 1
= arctan(2t) + ln = − ln 2
2 1 2 t2 1 8 2
2 2

(e)
Z 1
1 π
√ dt =
0 1−t 2 2

(f)
5 5
√
Z
1
√ dt = 2 t − 1 =4
1 t−1 1

14
 (g)
Z 2 Z 2 √
t 1 8
√ dt = t−1+ √ dt =
1 t−1 1 t−1 3
√
(h) By the change of variable s = t − 1, i.e. t = s2 + 1:
Z +∞ Z +∞
1 1
√ dt = 2 ds = π
1 t t−1 0 1 + s2
√
(i) By the change of variable s = t, i.e,. t = s2 ,
Z +∞ Z +∞
1 1
√ dt = 2 ds = π
0 t(1 + t) 0 1 + s2
√
17. (a) By the change of variable s = t, i.e., t = s2 ,
Z +∞ √ Z +∞
e− t dt = 2se−s ds = 2Γ(2) = 2
0 0

(b)
Z 1 Z 1 Z 1
3 3 3 3
ln t 2 dt = (t)0 ln t dt = − 1 dt = −
0 2 0 2 0 2

(c)
Z +∞ Z +∞ Z +∞
ln t 1 −2 0 1 1 1
3
dt = ln t(−t ) dt = 3
dt =
1 t 2 1 2 1 t 4

(d) Integrating by parts twice,

Z +∞
5
t2 e−t dt =
1 e

(e) By the change of variable s = et , i.e. t = ln s:

Z +∞ Z +∞ Z +∞
1 2et 1 π π π
dt = 2t
dt = 2 2
dt = 2( − ) =
0 cosh t 0 1 + e 1 1 + s 2 4 2

(f)
+∞ ∞
arctan2 t π2
Z
arctan t
dt = =
0 1 + t2 2 0 8

(g)
Z +∞ ∞ Z +∞
arctan t arctan t 1 1 π 1 π 1
dt = − + dt = + − =
1 t3 2t2 1 2 1 t2 (1 + t2 ) 8 2 8 2

18. (a) The integral converges if α < 1 and diverges otherwise.

(b) The integral converges if α > 1 and diverges otherwise.
(c) The integral converges if −1 < α < 0 and diverges otherwise.
(d) The integral converges if −1 < α and diverges otherwise.

15
 (e) The integral converges if α < 1 and diverges otherwise.
(f) The integral converges if α > 1 and diverges otherwise.
19. (a) True. Without loss of generality, assume that f is strictly increasing and let x1 , x2 ∈ R
be such that x1 < x2 . Then f (x1 ) < f (x2 ) and so f (x1 ) 6= f (x2 ).
(b) False. Take for example f (x) = 2[x] − x + 1.
(c) False. Take for example f (x) = f −1 (x) = x.
(d) True. Let y ∈ R. There exists x ∈ R such that f (x) = y. Then we have f −1 (−y) =
f −1 (−f (x)) = f −1 (f (−x)) = −x = −f −1 (y).
(e) False. Take for example f (x) = x and g(x) = x2 that satisfy (f ◦g)(x) = x2 = (g ◦f )(x)
with f 6= g.
(f) True. Let x1 , x2 ∈ R be such that x1 6= x2 . By the injectivity of g, we have g(x1 ) 6=
g(x2 ) and so f (g(x1 )) 6= f (g(x2 )) by the injectivity of f .
(g) True. Let x1 , x2 ∈ R be such that f (x1 ) = f (x2 ). Then we have f (f (x1 )) = f (f (x2 )).
Since f ◦ f is injective, we conclude that x1 = x2 .
(h) True. Let x1 , x2 ∈ R be such that g(x1 ) = g(x2 ). Then we have f (g(x1 )) = f (g(x2 )).
Since f ◦ g is injective, we conclude that x1 = x2 .
(i) False. Take for example f (x) = x2 and g(x) = ex defined from R to R. f is not injective
but (f ◦ g)(x) = e2x is injective.
(j) True. Let y ∈ R, then there exists x ∈ R such that (f ◦ g)(x) = y. Then there exists
z = g(x) such that f (z) = y, and therefore f is surjective.
(k) False. Take for example f (x) = x and g(x) = −x.
(l) False. Take for example A = {0, 1}, B = {0, 2}and f (0) = a, f (1) = f (2) = b. Then
we have f ([A ∩ B]) = {a} =
6 {a, b} = f ([A]) ∩ f ([B]).
1
20. (a) True. The result follows from the Leibniz critrion. The sequence ak = ln(k+1) is
decreasing and tends to 0.
ak+1 k+1 1
(b) True. We use d’Alembert’s criterion: ak = 2k+1 → 2 < 1.
Pn 1
(c) False. We have seen in the course that if 0 < α ≤ 1, then the series k=1 kα diverges.
p √
k
k
(d) True. We use Cauchy’s criterion: k |ak | = ln(k) → 0 < 1.

21. (a) False. The statement is false if, for example, lim f (x) = l0 6= l.
x→−∞

(b) True. For  = 1, there exists M > 0 such that if x ≥ M then |f (x) − l| < 1; in other
words, l − 1 < f (x) < l + 1 for all x ∈ [M, +∞[.
(c) True. As above, there exists M > 0 such that f is bounded on [M, +∞[. Since f is
even, it is also bounded on ] − ∞, −M ]. Since f is continuous on the closed bounded
interval [−M, M ], f also bounded on this interval. Hence f bounded on all of R.

1
 if x > 0
(d) False. Take, for example, f (x) = 0 if x = 0 .

−1 if x < 0


(e) True. Suppose that there were x ∈ R such that f (x) = l0 > l. Since lim f (x) = l,
x→+∞
there exists C > 0 such that |f (x) − l| < l0 − l, and so f (x) < l0 , for all x > C. But
since f is periodic, there exists T > 0 such that x + T > C et f (x + T ) = l0 . This is a
contradiction.
(f) True. This is the definition of a horizontal asymptote.

16
  [x]+1
 [x]
 if [x] is odd
[x]+2
(g) False. Take, for example, f (x) = if [x] is even and [x] > 0 .
 [x]
2 if [x] = 0

We have lim f (x) = 1, f (x) > 1 for all x ≥ 0, but there exist no M such that f is
x→+∞
monotone on ]M, +∞[.
(
[x]
[x]+1 if [x] 6= 1
(h) False. Take, for example, f (x) = . One has lim f (x) = 1, but f
0 if [x] = −1 x→+∞

is discontinuous at every x ∈ N.

22. (a) True. We have lim f (h)

h = 0, so for  = 1 there exists δ > 0 such that f (h)
h < 1 if
h→0
|h| < δ. In other words, |f (h)| < 1 · |h| if h is sufficiently small.
(b) False. Take, for example, f (h) = h. We have f (h) ≤ 1 · |h|, but on the other hand
lim f (h)
h = 1.
h→0

(c) False. Take, for example, f (h) = h. We have f (h) = O(h), but f (h) 6= O(h2 ).
g(h)
(d) True. There exists C > 0 such that |g(h)| ≤ C|hq | for h small. Thus hp <
q−p p
C|h | −→ 0 and it follows that g(h) = o(h ). We conclude by recalling that
h→0
p p p
o(h ) + o(h ) = o(h ).
(e) True. There exists C > 0 such that g(h)
hq < C for h small. Thus f (h)g(h)
hp+q = fh(h)
p
g(h)
hq is a
product of a function which tends toward 0 and a bounded function and so it converges
to 0.
(
x2 if x ∈ Q
(f) False. Take, for example, f (x) = . This function is continuous only
0 if x ∈ R\Q
at x = 0 and it is differentiable at this point since lim f (x)
x = 0.
x→0
(g) False. Take, for example, f (x) = |x| which is not differentiable 0. The left and right
derivatives exist but are different at this point.
(
x2 sin x1 if x 6= 0
(h) False. Take, for example, f (x) = . This function is differentiable
0 if x = 0
on ] − 1, 1[ but its derivative is not continuous at 0.
(
x if x ∈ Q
(i) False. Take, for example, f (x) = 2
. We have f 0 (0) = 1 but this
x + x if x ∈ R\Q
function not monotone on any neighbourhood of 0.
(j) True. This follows from the fact that C a (R) ⊂ C b (R) if a ≥ b.
√
(k) False. Take f (x) = x. We have g(x) = x2 = |x| which is not differentiable at 0.
(l) False. We have f (0) = 1. Thus (f −1 )0 (1) = 1
f 0 (0) = 1
1+e0 = 12 .
(m) True. We have f 0 (1) = 2 − 2 = 0. Hence (f ◦ f )0 (1) = f 0 (f (1)) · f 0 (1) = 0.
23. (a) True. This is the statement of l’Hospital’s rule.
f 0 (x)
(b) False. Take, for example, f (x) = x + sin x and g(x) = x. We have g 0 (x) = 1 + cos x
which has no limit at infinity. However, lim f (x) = lim 1 + sin x
= 1.
x→+∞ g(x) x→+∞ x

(c) False. Take the counterexample for the preceding question.

(d) True. f and g are differentiable on [x, y], so by the mean value theorem, there exists
0
c ∈]x, y[ such that fg0 (c)
(c)
= fg(y)−g(x)
(y)−f (x)
= 1.

17
 (e) True. Since g is differentiable on R, the function sin g(x) is differentiable on R. If
g(a) = 0, then sin g(a) = 0 and we may apply l’Hospital’s rule which guarantees the
existence of the limit. If g(a) 6= 0, then by continuity, the limit is simply sing(a)
g(a)
.
(f) False. Take, for example, g(x) = x and a = 1. We have g(a) = 1 6= 0 (note that in this
case l’Hospital’s rule does not apply). By continuity, the limit is sinh
1
1
6= cosh 1.

24. (a) False. Since 7 is odd, f admits an inflexion point at a.

(b) True. Let f (x) = a0 + a1 x + · · · + an xn + o(xn ) be the Taylor expansion of f to order
n at 0. We then have f (−x) = a0 − a1 x + · · · + (−1)n an xn + o(xn ). Since f is odd
and the Taylor expansion is unique, we conclude that a2m = 0 for all 0 ≤ 2m ≤ n. As
(k)
ak = f k!(0) , we obtain the result.
(c) True. This was shown in class.

18