Scribd
Upload
138 views

Sol 5

This document contains solutions to problems from Math 112 at the end of Chapter 4. It includes solutions to problems on continuity of functions, properties of continuous functions on compact sets, addition and scalar multiplication maps being continuous, properties of continuous functions on connected sets, the intermediate value theorem, uniform continuity, and Riemann integration. Detailed proofs and explanations are provided for each problem.

Uploaded by

InstantRamen
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
138 views

Sol 5

This document contains solutions to problems from Math 112 at the end of Chapter 4. It includes solutions to problems on continuity of functions, properties of continuous functions on compact sets, addition and scalar multiplication maps being continuous, properties of continuous functions on connected sets, the intermediate value theorem, uniform continuity, and Riemann integration. Detailed proofs and explanations are provided for each problem.

Uploaded by

InstantRamen
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Problem Set 5 Solution Set

Anthony Varilly
(adapted from Nitin Saksenas LATEXfile)
Math 112, Spring 2002

1. End of Chapter 4, Exercise 6b. Let f : R R be a bounded function. Prove that f is


continuous if and only if the graph of f is a closed subset of R2 . What if f is unbounded?
Solution. (=) Assume f is continuous. Let Gf = {(x, f (x)) | x R} denote the graph of f .
Since f is continuous, for each convergent subsequence xk x0 , we have f (xk ) f (x). Thus
every sequence of the form (xn , f (xn )) Gf converges to (x, f (x)) Gf . But this means
that Gf is closed.
(=) Assume that Gf is closed. Since f is bounded, say |f (x)| M , every sequence
(xn , f (xn )) in Gf such that xn x is contained in ({xn } {x}) [M, M ]. This is the
product of two compact spaces and so it is compact in R2 . Hence (xn , f (xn )) has a convergent
subsequence (xnk , f (xnk )) in this space. Since the limit point lies in Gf , it must be of the
form (x, f (x)). Hence f (xnk ) f (x) and so f is continuous.
If f is unbounded then Gf could be closed but f could be discontinuous. Consider the function
(
1/x x > 0
f (x) =
0
x0
The graph of this function is closed (its complement is openalso, it is the union of two graphs
that are closed). But the function is discontinuous at zero.
2. End of Chapter 4, Exercise 7. Consider a compact set B Rn and let f : B Rm be
continuous and one-to-one. Then prove that f 1 : f (B) B is continuous. Show by
example that this may fail if B is connected but not compact.
Solution. Consider the closed set C B Rn . Since B is closed and bounded and C is a
closed subset of B, C is also bounded and therefore compact by the HeineBorel theorem.
Since f is continuous, it follows that f (C) is compact and thus closed. We then claim that
(f 1 )1 (C) = f (C). Indeed, since f is one-to-one, it maps B bijectively onto its image f (B).
One can then easily show inclusions both ways.
Hence given a closed set C B, (f 1 )1 (C) is closed, so f 1 is continuous.
To show that this may fail if B is connected but not compact, consider f : [0, 2) R2
given by f (t) = (sin t, cos t). Observe that f ([0, 2)) equals the unit circle S R2 . (Also f is
one-to-one and continuous.) But the preimage of f 1 , which equals f , maps an open set to
a closed set so f 1 cannot be continuous.
1

3. End of Chapter 4, Exercise 8. Define maps s : Rn Rn Rn and m : R Rn Rm as


addition and scalar multiplication maps defined by s(x, y) = x + y and m(, x) = x. Show
that these mappings are continuous.
Solution. Consider the continuous functions s1 (x, y) = x and s2 (x, y) = y. Then s = s1 + s2
and by Corollary 4.3.3, s is continuous. Now consider the continuous functions m1 (, x) =
and m2 (, x) = x. Then m = m1 m2 and by Corollary 4.3.3, m is continuous.
4. End of Chapter 4, Exercise 18. Let A M be connected and let f : A R be continuous
with f (x) 6= 0 for all x A. Show that f (x) > 0 for all x A or else f (x) < 0 for all x A.
Solution I. Suppose there exist points a, b of A for which f (a) < 0 and f (b) > 0. Then by
the Intermediate Value Theorem there exists a point c (a, b) such that f (c) = 0, contrary
to the hypothesis on f . Thus either f (x) > 0 for all x A or f (x) < 0 for all x A.
Solution II. Suppose there exist points a, b of A for which f (a) < 0 and f (b) > 0. Define the
sets
U = {x M | f (x) < 0}
V = {x M | f (x) > 0}.
We are assuming U and V are non-empty. Note that both U and V are open, since, for
example, U = f 1 ((, 0)), i.e., it is the pre-image of an open set under a continuous
mapping. Then U and V form a separation of A, contrary to the hypothesis that A is
connected. Hence one of the sets U or V must be empty in fact.
5. End of Chapter 4, Exercise 29. Let f : R R satisfy |f (x) f (y)| |x y|2 . Prove that f
is a constant.
|f (x) f (y)|
Solution. |f (x) f (y)| |x y|2
|x y| (since |x y| is non-negative)
|x y|



f (y) f (x)
f (y) f (x)



limyx |y x| |f 0 (x)|

|y x|. It follows that limyx
yx
yx
0 f 0 (x) = 0 f is constant.
6. End of Chapter 4, Exercise 30a. Let f : [0, ) R, f (x) =
continuous.

x. Prove that f is uniformly

Solution I. Well prove f is uniformly continuous in [0, 1/4] and on (1/4, +). The former
follows from the Uniform Continuity theorem since f is continuous and [0, 1/4] is compact.

For the latter, notice f 0 (x) = 1/2 x < 1 for x > 1/4. Since f has bounded derivative (i.e., it
is Lipshitz), it is uniformly continuous in (1/4, +).

Solution II. Let  > 0 be given and choose = 2 . We want to show that for every pair of

points x, y such that |x y| < ,


|f (x) f (y)| = | x y| = | y x| < . Observe that

|x y| < y < x + y < x + . Thus

|f (x) f (y)| = | y x| < | x + x| < | x + x| = = 


as desired. Note that the inequality

x+y <

x+

y is valid for x, y [0, +).

7. End Chapter 4, Exercise 34. Assuming that the temperature on the surface of the earth is a
continuous function, prove that on any great circle of the earth there are two antipodal points
with the same temperature.
Solution. Pick a great circle on the earth and a point p on that circle. Let T () be the
temperature at rads, measured counter-clockwise from p. Consider the function S() =
T () T ( + ). S is continuous since T is. It follows that S(0) = T (0) T () and
S() = T () T (2) = T () T (0), S(0) = S(). From the Intermediate Value Theorem,
there exists c (0, ) such that S(c) = 0, which implies that T (c) = T (c + ). Thus the
temperature at two antipodal points is the same.
8. End of Chapter 4, Exercise 42. For x > 0 define L(x) =
this definition:

Rx
1

(1/t)dt. Prove the following using

(a) L is increasing in x. From part (c), we have L0 (x) = 1/x which is positive when x is
positive. Hence L is increasing.
(b) L(xy) = L(x) + L(y). Fix x > 0 and let f (z) = L(xz). Since z 7 xz is differentiable at
every z > 0 with derivative x, we can use part (c) and the chain rule to get
f 0 (z) = L0 (xz) x = (1/xz) x = 1/z = L0 (z)
Since (f L)0 = 0, f L is constant. In particular,
(f L)(y) = (f L)(1) = f (1) L(1) = L(x) 0
since L(1) = 0 from part (d). Thus L(x) = f (y) L(y) = L(xy) L(y) which implies
L(xy) = L(x) + L(y) as desired.
(c) L0 (x) = 1/x. For x > 0, f : [1, x] R (and f : [x, 1] R in the case 0 < x < 1) given by
f (t) = 1/t is continuous and thus satisfies the hypothesis of the Fundamental Theorem
of Calculus. Using the version derived in class, we have L0 (x) = F 0 (x) = f (x) = 1/x.
R1
(d) L(1) = 0. By definition L(1) = 1 (1/t)dt = 0.
(e) Properties (c) and (d) uniquely determine L. What is L? Suppose there are two
functions L(x) and M (x) such that L0 (x) = M 0 (x) = 1/x and L(1) = M (1) = 0. Then
(L M )0 = 0 (L M ) is a constant. Thus (L M )(1) = 0 (L M )(x) = 0 from
which it follows that L(x) = M (x). Therefore properties (c) and (d) uniquely determine
L. L(x) = log(x).

Remark. Many people had trouble with this exercise. In analysis L(x) is the definition of
the natural logarithm. So we were not allowed to use the known properties of ln to prove
the above statements. Also, part (b) required a clever proof since we have not yet proven the
change of variables formula for integration.
9. End of Chapter 4, Exercise 44. Let f : [0, 1] R be Riemann integrableR and suppose for
1
every a, b with 0 a < b 1 there is a c, a < c < b, with f (c) = 0. Prove 0 f = 0. Must f
be zero? What if f is continuous?
Solution. For any partition
f (xk ) = 0. Consider

Sn

of [0, 1], we know that there exist xk Ik for which

k=1 Ik

s0 =

n
X

f (xk )|Ik |

k=1

Because f (xk ) = 0 for all k, we know that s0 = 0. We also know that inf{f (x) | x Ik }
R1
R1
f (xk ) sup{f (x) | x Ik }. Consequently, 0 f s0 0 f . But the definition of integrability
R1
R1
implies 0 f (x)dx = 0 f (x)dx and thus
Z

f (x)dx = s0 = 0.

The function f need not be zero. Consider the following function f : [0, 1] R:
1 x = 1/2
0 x=
6 1/2

f (x) =

This function satisfies the hypotheses of this exercise and integrates to zero (the lower integral
is clearly 0 and the upper integral goes to 0 as the partitions of the interval get finer). If f is
continuous however, it must be zero. If f were not identically zero, then in order to satisfy the
condition that for every a, b with 0 a < b 1 there is a c such that a < c < b, f would have
to have an infinite number of jump discontinuities which would make it discontinuous.
Remark. Many people tried using the example f : [0, 1] R
(
1 xQ
f (x) =
0 x
/Q
This function isR not Riemann integrable. It is only integrable in the sence of Lesbesgue. So
when we write f = 0 we mean the Lebesgue integral of the function is zero. Look at your
class notes for details.

You might also like