MATH1010 University Mathematics

Additional exercise 5 Solution

Mean value theorem
x √
1. Let f (x) = 1 + − 1 + x. Then f (0) = 0 and
2
1 1
f 0 (x) = − √ >0
2 2 1+x
x √
for x > 0. Thus 1 + > 1 + x for x > 0. On the other hand, let
2
√ x x2
g(x) = 1 + x − 1 − + . Then g(0) = 0 and
2 8
1 1 x
g 0 (x) = √ − +
2 1+x 2 4
1 1 x
> x − +
2(1 + 2 ) 2 4
4 − 2(2 + x) + x(2 + x)
=
4(2 + x)
2
x
=
4(2 + x)
> 0
x x2 √
for x > 0. Thus 1 + − < 1 + x for x > 0.
2 8
2. Let f (x) = sin x tan x−2 ln sec x. Then f (0) = 0 and for any x ∈ (0, π2 ),

2 sec x tan x
f 0 (x) = cos x tan x + sin x sec2 x −
sec x
sin x 2 sin x
= sin x + 2
−
cos x cos x
2
sin x(cos x + 1 − 2 cos x)
=
cos2 x
sin x(1 − cos x)2
=
cos2 x
> 0

for x ∈ (0, π2 ). Thus sin x tan x > 2 ln sec x for x ∈ (0, π2 ).

1
 x3
3. Let f (x) = tan−1 x − x + . Then f (0) = 0 and
3
1 1 − (1 + x2 ) + x2 (1 + x2 ) x4
f 0 (x) = − 1 + x 2
= = >0
1 + x2 1 + x2 1 + x2
x3
for x > 0. Thus f (x) > 0 for x > 0. Therefore tan−1 x > x − for
3
x > 0.
θ2
4. Let f (θ) = cos θ − 1 + . Then
2
f 0 (θ) = θ − sin θ
f 00 (θ) = 1 − cos θ
Now f 0 (0) = 0 and f 00 (θ) > 0 for θ > 0 implies that f 0 (θ) > 0 for θ > 0.
Combining this with f (0) = 0, we have f (θ) > 0 for θ > 0. Thus
θ2 θ2 θ4
cos θ > 1− for θ > 0. On the other hand, let g(θ) = 1− + −cos θ.
2 2 24
Then
θ3
g 0 (θ) = −θ + + sin θ
6
θ2
g 00 (θ) = −1 + + cos θ
2
θ2
Now g 0 (0) = 0 and we have proved that g 00 (θ) = cos θ − 1 + > 0 for
2
any θ > 0. Thus g 0 (θ) > 0 for any θ > 0. Combining this with g(0) = 0,
θ2 θ4
we see that g(θ) > 0 for any θ > 0. Therefore 1 − + > cos θ for
2 24
θ > 0.
a1 t2 a2 t3 an tn+1
5. Let f (t) = + + ··· + . Then
2 3 n+1
f 0 (t) = a1 t + a2 t2 + · · · + an tn .
By mean value theorem, there exists 0 < x < 1 such that
f (1) − f (0)
f 0 (x) =
1−0
a1 a2 an
a1 x + a2 x 2 + · · · + an x n = + + ··· + .
2 3 n+1

2
6. Suppose f (a) = f (b) = 0.

(a) Let h(x) = ex f (x) which is continuous on [a, b] and differentiable

on (a, b). Then h(a) = h(b) = 0 and h0 (x) = ex (f 0 (x) + f (x)). By
Rolle’s theorem, there exists a < ξ < b such that

h0 (ξ) = eξ (f 0 (ξ) + f (ξ)) = 0

which implies f 0 (ξ) + f (ξ) = 0 since eξ > 0.

(b) Let p(x) = eg(x) f (x) which is continuous on [a, b] and differen-
tiable on (a, b). Then p(a) = p(b) = 0 and p0 (x) = eg(x) (f 0 (x) +
g 0 (x)f (x)). By Rolle’s theorem, there exists a < η < b such that

p0 (η) = eg(η) (f 0 (η) + g 0 (η)f (η)) = 0

which implies f 0 (η) + g 0 (η)f (η) = 0 since eg(η) > 0.

7. Let g(x) = e−x f (x) − f (0). Then g(0) = 0 and

g 0 (x) = −e−x f (x) + e−x f 0 (x) ≤ −e−x f (x) + e−x f (x) = 0

for any x > 0. Thus g(x) = e−x f (x) − f (0) ≤ 0 for any x > 0 which
implies f (x) ≤ f (0)ex for any x > 0.

8. (a) Note that

(a − b)(a − c)
F (a) = f (a) − f (a) = 0
(a − b)(a − c)
(b − c)(b − a)
F (b) = f (b) − f (b) = 0.
(b − c)(b − a)
By Rolle’s theorem, there exists ζ ∈ (a, b) such that F 0 (ζ) = 0.
(b) Note that

F 0 (x)
(2x − (b + c))f (a) (2x − (c + a))f (b) (2x − (a + b))f (c)
= f 0 (x) − − −
(a − b)(a − c) (b − c)(b − a) (c − b)(c − a)
F 00 (x)
2f (a) 2f (b) 2f (c)
= f 00 (x) − − −
(a − b)(a − c) (b − c)(b − a) (c − b)(c − a)

3
 By (a), there exists ζ ∈ (a, b) such that F 0 (ζ) = 0. Similarly, there
exists ξ ∈ (b, c) such that F 0 (ξ) = 0. By applying Rolle’s theorem
to F 0 (x), there exists a < ζ < η < ξ < b such that

F 00 (η) = 0
2f (a) 2f (b) 2f (c)
f 00 (η) − − − =0
(a − b)(a − c) (b − c)(b − a) (c − b)(c − a)
f 00 (η) f (a) f (b) f (c)
= + +
2 (a − b)(a − c) (b − c)(b − a) (c − b)(c − a)

9. (a) Let f (x) = x ln x for x > 0. By mean value theorem, there exists
ξ ∈ (a, b) such that
f (b) − f (a)
f 0 (ξ)
b−a
0
Since f (x) = 1 + ln x is strictly increasing, we have

f 0 (a) f 0 (ξ)
< < f 0 (b)
f (b) − f (a)
f 0 (a) < < f 0 (b)
b−a
b ln b − a ln a
1 + ln a < < 1 + ln b
b−a
(1 + ln a)(b − a) < b ln b − a ln a < (1 + ln b)(b − a)
1
(b) Let f (x) = xe x . Then
   
0 11 1 11
f (x) = e + xe − 2 = 1 −
x exx
x x

Applying mean value theorem to f (x) on ( 1b , a1 ), there exists a <

c < b such that
f ( a1 ) − f ( 1b ) 1
1 1 = f 0( )
a
−b c
ea eb
−
a
1 1
b
= (1 − c)ec
a
− b
a b
be − ae
= (1 − c)ec
b−a
aeb − bea = (b − a)(c − 1)ec

4
10. (a) For any x, by applying mean value theorem to f (x) on (x, x + 1),
there exists x < ξ < x + 1 such that
f (x + 1) − f (x)
f 0 (ξ) = = f (x + 1) − f (x)
x+1−x
Since f 0 (x) < 0 for any x, f 0 (x) is strictly decreasing. Thus f 0 (x +
1) < f 0 (ξ) < f 0 (x) which implies f 0 (x + 1) < f (x + 1) − f (x) <
f 0 (x).
(b) By (a),

f 0 (1) < f (1) − f (0) < f 0 (0)

f 0 (2) < f (2) − f (1) < f 0 (1)
f 0 (3) < f (3) − f (2) < f 0 (2)

Adding up the above inequalities, we have

f 0 (1) + f 0 (2) + f 0 (3) < f (3) − f (0) < f 0 (0) + f 0 (1) + f 0 (2).

11. Observe that at least one of the following holds. Either λ lies between
f (b) − f (a) f (b) − f (a)
f 0 (a) and , or λ lies between f 0 (b) and . Sup-
b−a b−a
f (b) − f (a)
pose λ lies between f 0 (a) and . Define
b−a

f 0 (a) if t = a
fa (t) = f (t) − f (a)
 if t 6= a
t−a
Then fa (t) is continuous on [a, b] since

f (t) − f (a)
lim fa (t) = lim = f 0 (a) = fa (a).
t→0 t→0 t−a
Also fa (t) is differentiable at any t ∈ (a, b) because f (x) is differentiable
at any x ∈ (a, b). Since λ lies between fa (a) = f 0 (a) and fa (b) =
f (b) − f (a)
, by applying intermediate value theorem to fa (t) on [a, b],
b−a
f (η) − f (a)
there exists η ∈ [a, b] such that fa (η) = = λ. Applying
η−a

5
 mean value theorem to f (x) on [a, η], there exists ξ ∈ (a, η) such that
f (η) − f (a)
f 0 (ξ) = = λ.
η−a
f (b) − f (a)
Suppose λ lies between f 0 (b) and . Define
b−a
(
f 0 (b) if t = b
fb (t) = f (b)−f (t)
b−t
if t 6= b
The rest of the argument is more or less the same.
12. (a) (i)
x x x y y y
x= + < + < + = y.
p q p q p q
(ii) Applying mean value theorem to f (x) on (x, z), there exists
x < ξ < z such that
f (z) − f (x)
f 0 (ξ) = .
z−x
Now  
x y x x y−x
z−x= + − + = .
p q p q q
Hence
q(f (z) − f (x))
f 0 (ξ) = .
y−x
(iii) Applying mean value theorem to f (x) on (z, y), there exists
z < η < y such that
p(f (y) − f (z))
f 0 (η) = .
y−x
Since f 00 (t) ≥ 0 for any t and ξ < η, we have f 0 (ξ) ≤ f 0 (η)
and therefore
q(f (z) − f (x)) p(f (y) − f (z))
≤
y−x y−x
(p + q)f (z) ≤ qf (x) + pf (y)
qf (x) pf (y)
f (z) ≤ +
p+q p+q
f (x) f (y)
= +
p q

6
 (b) (i) Without loss of generality, we may assume p ln a < q ln b. Take
x = p ln a, y = q ln b and
x y
z= + = ln a + ln b = ln(ab).
p q

By (a)(iii), we have

f (p ln a) f (q ln b)
f (ln(ab)) ≤ + .
p q

(ii) Let f (t) = et . Then f 00 (t) = et > 0 for any t ∈ R. By (b)(i),

we have

ln(ab) ep ln a eq ln b
e ≤ +
p q
ap b q
ab ≤ + .
p q

13. (a) Consider the function g(x) defined by

(
f (x) if x > 0
g(x) = .
0 if x = 0

Then lim+ g(x) = lim+ f (x) = 0 = g(0). Thus g(x) is continuous

x→0 x→0
on [0, +∞). Moreover, g 0 (x) = f 0 (x) > 0 for any x > 0. For any
x > 0, applying mean value theorem to g(x) on [0, x], there exist
0 < ξ < x such that

g(x) − g(0)
= g 0 (ξ)
x−0
f (x) − 0
= f 0 (ξ)
x
which implies f (x) = f 0 (ξ)x > 0.
(b) Consider f (x) = (1 + x) ln(1 + x) − x ln x. Then

f 0 (x) = ln(1 + x) + 1 − ln x − 1 = ln(1 + x) − ln x > 0

7
 for any x > 0 and

lim f (x) = lim (1 + x) ln(1 + x) − x ln x

x→0+ x→0+
= − lim+ x ln x
x→0
ln( 1t )
= − lim
t→+∞ t
ln t
= lim
t→+∞ t
= 0.

By (a), we have f (x) = (1 + x) ln(1 + x) − x ln x > 0 for any x > 0.

ln(1 + at )
(c) Let a > 0 and g(t) = . Then for any t > 0,
t
t
0
· at ln a − ln(1 + at )
1+at
g (t) =
t2
a ln(a ) − (1 + at ) ln(1 + at )
t t
=
t2 (1 + at )
< 0

where the last inequality follows from (b) since at > 0 for any
t > 0.
v
(d) Let a = . Note that g(x) is continuous on [p, q] and differentiable
u
on (p, q). By the mean value theorem, there exists ξ ∈ (p, q) such
that
g(q) − g(p)
= g 0 (ξ) < 0
q−p
ln(1 + aq ) ln(1 + ap )
− < 0
q p
1 1
ln(1 + aq ) q < ln(1 + ap ) p
1 1
(1 + aq ) q < (1 + ap ) p
1 1
vq q vp p
 
1+ q < 1+ p
u u
1 1
(uq + v q ) q < (up + v p ) p .