MATH 3162 Homework Assignment 2 Solutions

4.4.13(a): Assume that f : A → R is uniformly continuous on A and that

(xn ) is a Cauchy sequence in A. Choose any  > 0. By definition of uniform
continuity, there exists δ > 0 so that for any a, b ∈ A with |a − b| < δ, it is true
that |f (a)−f (b)| < . Since (xn ) is Cauchy, there exists N so that for n, m > N ,
|xn − xm | < δ. Therefore, |f (xn ) − f (xm )| < . But we have then shown that
for any  > 0, there exists N so that n, m > N implies |f (xn ) − f (xm )| < , and
so by definition (f (xn )) is Cauchy.

4.5.2(a): There are many examples of such objects. For instance, define f (x) =
sin x on (0, 2π). Clearly f is continuous on (0, 2π), and f ((0, 2π)) = [−1, 1].

4.5.2(b): This is impossible; any closed interval is compact, and so f ([a, b])
must be compact, but no open interval is compact.


1
4.5.2(c): There are many examples. For instance, define f (x) = 1−x 2 on

(−1, 1). Clearly f is continuous on (−1, 1), and f (−1, 1) = [1, ∞), which is
closed and unbounded, but not R.

4.5.3: Assume that f is increasing on [a, b] and satisfies the Intermediate Value
Property. Choose any c ∈ [a, b]; we want to show that f is continuous at c.
Start with the case c ∈ (a, b), and choose any  > 0. We claim that we can
find m, M so that m < c < M and f (m) > f (c) − , f (M ) < f (c) + . If
f (a) ≥ f (c) − 0.5, then just define m = a. If f (a) < f (c) − 0.5, then we can
use the Intermediate Value Theorem (with L = f (c) − 0.5) to find m ∈ (a, c)
for which f (m) = f (c) − 0.5 > f (c) − . Similarly, if f (b) ≤ f (c) + 0.5, then
just define M = b. If f (b) > f (c) + 0.5, then we can use the Intermediate
Value Theorem (with L = f (c) + 0.5) to find M ∈ (c, b) for which f (M ) =
f (c)+0.5 < f (c)+. Now, just define δ = min(c−m, M −c). If |x−c| < δ, then
x ∈ (c − δ, c + δ) ⊂ (m, M ), and so since f is increasing, f (m) ≤ f (x) ≤ f (M ),
meaning that f (c) −  < f (x) < f (c) +  =⇒ |f (x) − f (c)| < . Since  was
arbitrary, we have verified continuity of f at c.
We still have to deal with the cases c = a or c = b. If c = a, choose any
 > 0. We proceed just as above to find M > a so that f (M ) < f (a) + .
Then, just define δ = M − a. If |x − a| < δ, then x ∈ (a − δ, a + δ), and since
f is defined only on [a, b], in fact a ≤ x < M . Then, since f is increasing,
f (a) ≤ f (x) ≤ f (M ) =⇒ f (a) ≤ f (x) < f (a) +  =⇒ |f (x) − f (a)| < , and
again we verified continuity of f at a. The proof for c = b is almost identical.
 
f (x)−f (c)
5.2.6(a): By definition, we know that g 0 (c) = limx→c x−c . We claim
0
that also g (c) = limh→0 f (c+h)−f
h
(c)
To see this, take any sequence hn →
.
0 where hn 6= 0 for all n. Then define xn = c + hn ; clearly xn → c and
xn 6= c for all c. Therefore, by definition of limit, f (xxnn)−f
−c
(c)
→ g 0 (c). But
f (xn )−f (c) f (cn +h)−f (c) f (cn +h)−f (c)
xn −c = h , and so we know that h → g 0 (c). Since
f (c+h)−f (c)
(hn ) was arbitrary, we’ve shown that g 0 (c) = limh→0 h .

Extra problem 1: Assume that f is continuous and 1-1 on [a, b]. Since f
is 1-1, either f (a) < f (b) or f (a) > f (b). Let’s start with the case where
f (a) < f (b). We wish to prove that in this case, f is strictly increasing on [a, b].
Choose an arbitrary pair x, y with x < y, and assume for a contradiction that
f (x) ≥ f (y). Again, since f is 1-1, f (x) 6= f (y), and so f (x) > f (y). We claim
that either f (a) < f (y) or f (x) < f (b). Indeed, if neither of these were true,
then f (y) ≤ f (a) < f (b) ≤ f (x), meaning that f (y) < f (x), which is false. We
now have two cases.

Case 1: f (a) < f (y). Then f (a) < f (y) < f (x) and a < x < y (note
that a 6= x since f (a) 6= f (x)). So, by the Intermediate Value Theorem (with
L = f (y)), there exists c ∈ (a, x) so that f (c) = f (y). However, this contradicts
the fact that f is 1-1; c < x < y, so c 6= y, but f (c) = f (y).
Case 2: f (x) < f (b). Then f (y) < f (x) < f (b) and x < y < b (note
that y 6= b since f (y) 6= f (b)). So, by the Intermediate Value Theorem (with
L = f (x)), there exists c ∈ (y, b) so that f (c) = f (x). However, this contradicts
the fact that f is 1-1; x < y < c, so c 6= x, but f (c) = f (x).

In each case we have a contradiction, so our original assumption was wrong,

i.e. f (x) < f (y). Since x < y in [a, b] were arbitrary, we’ve shown that f
is strictly increasing on [a, b] when f (a) < f (b). The proof that f is strictly
decreasing when f (a) > f (b) is almost identical. (Or, you could replace f with
−f to say that this fact follows without loss of generality!)

Extra problem 2: Assume that f is continuous on [a, b], f (a) < 0 < f (b),
and define S = {x : f (x) ≤ 0}. Define c = sup S. Assume for a contradiction
that f (c) < 0. Then we define  = −f (c) (remember that f (c) is negative,
so  > 0), and by definition of continuity, there exists δ > 0 so that for every
x ∈ (c − δ, c + δ), f (x) ∈ (f (c) − , f (c) + ) = (2f (c), 0). In particular, this
means that f (c + 0.5δ) < 0. However, then by definition c + 0.5δ ∈ S, which
means that c is not an upper bound of S, a contradiction to definition of sup S.
Therefore, our original assumption was wrong, and f (c) is not negative.

2
Extra problem 3: Assume that f : R → R is differentiable on R and that
f 0 (c) > 0. Then by definition of derivative, if we define d(x) = f (x)−f
x−c
(c)
, then
limx→c d(x) = f 0 (c). Choose  = f 0 (c). Then by definition of convergence, there
exists δ > 0 so that for every x ∈ (c − δ, c + δ), d(x) ∈ (f 0 (c) − , f 0 (c) + ) =
(0, 2f 0 (c)). In particular, d(c − 0.5δ) > 0 and d(c + 0.5δ) > 0. Then,

f (c − 0.5δ) − f (c) f (c − 0.5δ) − f (c)

d(c − 0.5δ) = = > 0,
c − 0.5δ − c −0.5δ
which means that f (c − 0.5δ) − f (c) < 0 =⇒ f (c) > f (c − 0.5δ). Therefore, f (c)
is not a minimum value of f (x) on R. Similarly,

f (c + 0.5δ) − f (c) f (c + 0.5δ) − f (c)

d(c + 0.5δ) = = > 0,
c + 0.5δ − c 0.5δ
which means that f (c + 0.5δ) − f (c) > 0 =⇒ f (c) < f (c − 0.5δ). Therefore, f (c)
is not a maximum value of f (x) on R.

Extra problem 4: (a) Assume that f 0 (x) 6= 0 for all x ∈ [a, b]. Then there
cannot be values y, z ∈ [a, b] with 0 between f 0 (y) and f 0 (z); otherwise, by
Darboux’s Theorem, there would exist c between y and z with f 0 (c) = 0, a
contradiction. So, either f 0 (x) > 0 for all x ∈ [a, b] or f 0 (x) < 0 for all x ∈ [a, b].

(b) By part (a), either f 0 (x) > 0 for all x ∈ [a, b] or f 0 (x) < 0 for all x ∈ [a, b].
Assume that f 0 (x) > 0 for all x ∈ [a, b]. Then, for every y < z ∈ [a, b], by
the Mean Value Theorem, there exists c ∈ [y, z] with f 0 (c) = f (z)−f z−y
(y)
. Since
f 0 (c) > 0 by assumption, and z −y > 0, it must be true that f (z)−f (y) > 0, i.e.
f (y) < f (z). Since y < z were arbitrary, this shows that f is strictly increasing
on [a, b]. If instead it were the case that f 0 (x) < 0 for all x ∈ [a, b], then virtually
the same proof shows that f is strictly decreasing on [a, b].