MIDTERM 2 SOLUTIONS

MATH 478: ANALYSIS I

Problem 1. (a) (1 point) State the extreme value theorem.

Let M be a compact metric space and f : M → R a continuous function.

Then, f is bounded on M and attains its maximum and minimum values.

(b) (1 point) State the intermediate value theorem.

Let M be a connected metric space and f : M → R a continuous function.

If y ∈ R and f (a) > y and f (b) < y for some a, b ∈ M , then there exists x ∈ M
with y = f (x).

1
2

(c) (3 points) Prove that if f : [a, b] → R is continuous, f ([a, b]) is a closed bounded
interval.

Proof. Since [a, b] is compact and f : [a, b] → R continuous, the extreme value
theorem ensures f ([a, b]) has minimum and maximum values, say c and d, re-
spectively. Hence,
f ([a, b]) ⊂ [c, d].
Now let x1 , x2 ∈ [a, b] with f (x1 ) = c and f (x2 ) = d. Since [a, b] is connected, if
y ∈ [c, d], then there exists x ∈ [a, b] with f (x) = y. That is,

[c, d] ⊂ f ([a, b]).

Hence, f ([a, b]) = [c, d], and the proof is done. 

 3

Problem 2. (a) (1 point) State the inverse function theorem.

Let f : (a, b) → (c, d) be a differentiable surjection such that, for each x ∈

(a, b), f 0 (x) 6= 0. Then, the inverse function f −1 : (c, d) → (a, b) is differentiable
on (c, d) and
1
(f −1 )0 (y) = 0 −1
f (f (y))
for each y ∈ (c, d).

(b) (3 points) Let q ∈ N. Prove using the inverse function theorem or the
1
definition of the derivative that f (x) = x q is differentiable on (0, ∞) and that
1 1
f 0 (x) = x q −1 .
q
d n
(You are allowed to use the fact that dx
x = nxn−1 for natural numbers n.)

Proof. We can use the inverse function theorem. Let g : (0, ∞) → (0, ∞) be
given by g(y) = y q and note f = g −1 . We know already that

g 0 (y) = qy q−1 6= 0

for each y > 0. Hence, by the inverse function theorem, f is differentiable and
1 1 1 1
f 0 (x) = = 1 = x q −1 .
g 0 (f (x)) q(x )q−1
q q


4

Problem 3. (a) (1 point) State Taylor’s theorem with the Lagrange form of the
remainder.

Let f : (a, b) → R be (n + 1)-times differentiable, let x0 , x ∈ (a, b). Then,

there exists t between x0 and x for which
n
X f (k) (x0 ) f (n+1) (t)
f (x) − (x − x0 )k = (x − x0 )n+1 .
k=0
k! (n + 1)!

(b) (3 points) Suppose that f : (a, b) → R is f is twice-differentiable and f 00 ≥ 0 on

(a, b). Show that f lies above each of its tangent lines.

Proof. Fix x0 , x ∈ (a, b). Then, there exists t between x0 and x for which

f 00 (t)
f (x) − f (x0 ) − f 0 (x0 )(x − x0 ) = (x − x0 )2 ≥ 0.
2
Hence,
f (x) ≥ f (x0 ) + f 0 (x0 )(x − x0 ).
In other words, the graph of f lies above its tangent line based at x0 for each
x0 ∈ (a, b). 
 5

Problem 4. (3 point) Let f : [a, b] → R be a bounded function. Prove that

sup L(f, P ) ≤ inf0 U (f, P 0 ),

P P

where the supremum and infimum are taken over partitions P and P 0 of [a, b].

Proof. Fix a partition P . For any partition P 0 , monotonicity yields

L(f, P ) ≤ L(f, P ∪ P 0 ) ≤ U (f, P ∪ P 0 ) ≤ U (f, P 0 ).

Hence L(f, P ) is a lower bound for {U (f, P 0 ) : partitions P 0 }, and hence

L(f, P ) ≤ inf0 U (f, P 0 ).

P

Since this holds for each P , inf P 0 U (f, P 0 ) is an upper bound for {L(f, P ) :
partitions P }, and so
sup L(f, P ) ≤ inf U (f, P ).
P P


6

Problem 5. (a) (1 point) State what it means for f : [a, b] → R to be Riemann

integrable over [a, b] with integral I.

For each  > 0, there exists δ > 0 such that for each partition pair P, T of
[a, b],
mesh P < δ implies |R(f, P, T ) − I| < .

(b) (3 points) Suppose a < c < b, let f : [a, b] → R is a Riemann integrable function
over [a, b], and suppose g : [a, b] → R satisfies f (x) = g(x) for each x 6= c. Prove
directly from the definition of the Riemann integral that g is integrable on
[a, b] and
Z b Z b
g(x) dx = f (x) dx.
a a

Proof. We first fix a partition pair P, T and bound

|R(g, P, T ) − R(f, P, T )|
in terms of the mesh of P . Let
P = {x0 , . . . , xn } and T = {t1 , . . . , tn }
as usual and let K be the set of all indices k for which tk = c. Note, K may
contain at most two indices (and exactly two indices only when c = xk for some
k ∈ {1, . . . , n − 1} and tk = tk+1 = xk ). We have
n
X n
X
|R(g, P, T ) − R(f, P, T )| = f (tk )(xk − xk−1 ) − g(tk )(xk − xk−1 )
k=1 k=1
Xn
= (f (tk ) − g(tk ))(xk − xk−1 )
k=1
X
= (f (c) − g(c))(xk − xk−1 )
k∈K
X
≤ |f (c) − g(c)|(xk − xk−1 )
k∈K
≤ 2(mesh P )|f (c) − g(c)|.
 7

Next, let Z b
I= f (x) dx
a

and fix  > 0. Since f is integrable over [a, b], there exists δ > 0 such that

mesh P < δ implies |R(f, P, T ) − I| < .
2
Hence if  

mesh P < min δ, ,
4|f (c) − g(c)|
we have
 
|R(g, P, T ) − I| ≤ |R(g, P, T ) − R(f, p, T )| + |R(f, p, T ) − I| < + = .
2 2