Suggested Problems from Chapter 5

Haram Kim

Exercises 5.1
Let f be defined for all real x, and suppose that

|f (x) − f (y)| ≤ (x − y)2

for all real x and y. Prove that f is constant.

Solution.
f (t)−f (x)
Define ϕ(t) = t−x for t ̸= x. By the given condition,

f (t) − f (x)
0 ≤ |ϕ(t)| = ≤ |t − x|.
t−x

Taking the limit t → x, then lim ϕ(t) = 0. The definition of differentiation implies that f is
t→x
differentiable at every point x in R and f ′ (x) = 0.
Now, pick arbitrary two real numbers x1 < x2 . Since f is differentiable in R, f is continuous on
[x1 , x2 ] and differentiable in (x1 , x2 ). By the mean value theorem, there is a point y ∈ (x1 , x2 )
at which
f (x2 ) − f (x1 ) = (x2 − x1 )f ′ (y) = 0.
This means that f takes same value for every x, which implies that f is constant.

1
Exercises 5.3
Suppose g is a real function on R1 , with bounded derivative (|g ′ | ≤ M ). Fix ε > 0, and define
f (x) = x + εg(x). Prove that f is one-to-one if ε is small enough. (A set of admissible values
of ε can be determined which depends only on M .)

Solution.
1
We first consider when M > 0 case. Fix 0 < ε < M . Consider two real numbers x1 , x2 and
without loss of generality, we can assume x1 < x2 . Then,

f (x2 ) − f (x1 ) = (x2 − x1 ) + ε(g(x2 ) − g(x1 )). (1)

Since g is differentiable in R1 , g is continuous on [x1 , x2 ] and differentiable on (x1 , x2 ). By the

mean value theorem, there exists θ ∈ (x1 , x2 ) satisfying g(x2 ) − g(x1 ) = g ′ (θ)(x2 − x1 ). Then,
equation (1) becomes
f (x2 ) − f (x1 ) = (x2 − x1 )(1 + εg ′ (θ)).
The condition given in the problem and the setting of ε imply that |1 + εg ′ (θ)| > 1 − εM > 0.
Hence, f (x1 ) ̸= f (x2 ), which means that f is one-to-one for small enough ε.

Now, consider the other case, M = 0. Then, g ′ (x) = 0 for all x ∈ R1 . By the mean value
theorem, g is constant and if so, f (x) = x + εg(x) becomes f (x) = x + (constant), which is
obviously one-to-one.

2
Exercises 5.4
If
C1 Cn−1 Cn
C0 + + ··· + + = 0,
2 n n+1
where C0 , ..., Cn are real constants, prove that the equation

C0 + C1 x + · · · + Cn−1 xn−1 + Cn xn = 0

has at least one real root between 0 and 1.

Solution.
Define the following function
C1 2 C2 3 Cn n+1
f (x) = C0 x + x + x + ··· + x .
2 3 n+1
Then, f (0) = 0, f (1) = 0 (Latter equation comes from the given condition.) Since f is polyno-
mial, f is differentiable in R1 . Therefore, f is continuous on [0, 1], differentiable in (0, 1). By
the mean value theorem, there exists θ ∈ (0, 1) such that f ′ (θ) = 0 but

f ′ (x) = C0 + C1 x + · · · + Cn−1 xn−1 + Cn xn

so that θ becomes a root of the equation we want to find.

3
Exercises 5.15
Suppose a ∈ R1 , f is a twice-differentiable real function on (a, ∞), and M0 , M1 , M2 are the least
upper bounds of |f (x)|, |f ′ (x)|, |f ′′ (x)|, respectively, on (a, ∞). Prove that M12 ≤ 4M0 M2 .
To show that M12 = 4M0 M2 can actually happen, take a = −1, define



 2x2 − 1 (−1 < x ≤ 0),

f (x) =
2
x − 1


 (0 ≤ x < ∞),
x2 + 1
and show that M0 = 1, M1 = 4, M2 = 4.
Does M12 ≤ 4M0 M2 hold for vector-valued functions too?

Solution.
Pick arbitrary x (x > a) and h > 0, and fix them. Since f is a twice-differentiable on (a, ∞),
we can say that f , f ′ are continuous on [x, x + 2h], and f ′′ exists in (x, x + 2h). Then, we can
apply the Taylor’s theorem, which implies that there exists cx ∈ (x, x + 2h) such that

f (x + 2h) = f (x) + 2hf ′ (x) + 2h2 f ′′ (cx ). (2)

With simple algebra,

1
f ′ (x) = (f (x + 2h) − f (x)) + hf ′′ (cx ).
2h
1 M0
|f ′ (x)| ≤ (|f (x + 2h)| + |f (x)|) + h|f ′′ (cx )| ≤ + hM2 . (3)
2h h
The second inequality in (3) is from the least upper bounds of |f (x)| and |f ′′ (x)| on whole
p
domain. Since h is arbitrary, we can take h = M0 /M2 to make

|f ′ (x)| ≤ 2 M0 M2 .
p

This inequality is true for every x > a, we can take the supremum on the both sides, making
√
a conclusion that M1 ≤ 2 M0 M2 . Each terms is positive so we can take square of each side.
Then,
M12 ≤ 4M0 M2 .

It is easy to check that the explicitly given function f (x) in the problem is twice-differentiable
with f ′ (0) = 0, f ′′ (0) = 4 and M0 = 1, M1 = 4, M2 = 4, which shows the case when equality
holds. (Be careful when trying to find the values of f ′ (0) and f ′′ (0).)

The key point in this proof is inequality (3). Now, we will prove (3) for vector-valued functions.
Then, the remaining part of proof is almost same.
Let f be a vector-valued function that is twice-differentiable on (a, ∞). Take same settings for
Mi for i = 0, 1, 2. Pick arbitrary β (β > a) and h > 0, and fix them. Now, define a new function

ϕ(x) = f (x) · f (β + 2h) − f (β) − 2hf ′ (β)

on x > a by taking (usual Euclidean) inner product of two vectors. For convenience, denote
z := f (β +2h)−f (β)−2hf ′ (β), and ϕ(x) = f (x)·z. Then, ϕ is a real valued function defined on
the real numbers x (x > a) and twice differentiable in (a, ∞) since ϕ is the linear combination

4
of each component of f , which is twice differentiable. Hence, we can apply the Taylor’s theorem
for ϕ on [β, β + 2h], saying that we can find γx ∈ (β, β + 2h) such that

ϕ(β + 2h) = ϕ(β) + 2hϕ′ (β) + 2h2 ϕ′′ (γx ).

|z|2 = (f (β + 2h) − f (β) − 2hf ′ (β)) · z

= ϕ(β + 2h) − ϕ(β) − 2hϕ′ (β)
= 2h2 ϕ′′ (γx )
= 2h2 f ′′ (γx ) · z
≤ 2h2 |f ′′ (γx )||z|

by the Cauchy-Schwarz inequality for the last inequality. Now, we obtain the inequality

|z| = |f (β + 2h) − f (β) − 2hf ′ (β)| ≤ 2h2 |f ′′ (γx )|

Then, using the property of absolute value, we can get

1
|f ′ (β)| ≤ |f (β + 2h)| + |f (β)| + h|f ′′ (γx )|,


(4)
2h
which is the same form of (3).

5
Exercises 5.17
Suppose f is a real, three times differentiable function on [−1, 1], such that

f (−1) = 0, f (0) = 0, f (1) = 1, f ′ (0) = 0.

Prove that f (3) (x) ≥ 3 for some x ∈ (−1, 1).

Note that equality holds for 12 (x3 + x2 ).

Solution. Since f is three times differentiable on [−1, 1], f , f ′ , and f ′′ are continuous on
[−1, 1]. Plus, f (3) (x) exists for every x ∈ (−1, 1). Therefore, we can apply the Taylor’s theorem
for α = 0, β = ±1, saying that we can find s ∈ (0, 1) and t ∈ (−1, 0) such that
1 1
f (1) = f (0) + f ′ (0)(1 − 0) + f ′′ (0)(1 − 0)2 + f (3) (s)(1 − 0)3 ,
2 6
′ 1 ′′ 1
f (−1) = f (0) + f (0)(−1 − 0) + f (0)(−1 − 0) + f (3) (t)(−1 − 0)3 .
2
2 6
Putting the value given and taking subtraction of two equations above, we obtain

f (3) (s) + f (3) (t) = 6. (5)

Hence, at least one of f (3) (s) and f (3) (t) is greater or equal to 3, or equation (5) cannot be
satisfied. In either case, we can conclude the existence of x ∈ (−1, 1) such that f (3) (x) ≥ 3.

It is easy to check that the function g(x) = 21 (x3 + x2 ) satisfies all the conditions given in
the problem and g (3) (x) = 3 for all real number x.