Math 317 HW #9 Solutions

1. Exercise 4.3.7. Assume h : R R is continuous on R and let K = {x : h(x) = 0}. Show that K is a closed set. Proof. Notice that, since h is dened on all of R, K c = {x : h(x) = 0}; showing that K c is open is equivalent to showing that K is closed. Of course, K c = h1 (R\{0}) = h1 ((, 0) (0, +)). Since (, 0) (0, +) is open, problem 5 below implies that K c is also open. 2. Exercise 4.3.9 (Contraction Mapping Theorem). Let f be a function dened on all of R, and assume there is a constant c such that 0 < c < 1 and |f (x) f (y )| c|x y | for all x, y R. (a) Show that f is continuous on R. Proof. Let > 0. Then whenever |x y | < , we have that |f (x) f (y )| c|x y | < c < since 0 < c < 1. Therefore, f is uniformly continuous on R, which implies that f is continuous on R. (b) Pick some point y1 R and construct the sequence (y1 , f (y1 ), f (f (y1 )), . . .). In general, if yn+1 = f (yn ), show that the resulting sequence (yn ) is a Cauchy sequence. Hence we may let y = lim yn . Proof. I claim that, if d = |y1 f (y1 )|, then |yn yn+1 | dcn1 . The proof is by induction: Base Case: Note that |y1 y2 | = |y1 f (y1 )| = d = dc0 . Inductive Step: Suppose |yk yk+1 | dck1 . Then |yk+1 yk+2 | = |f (yk ) f (yk+1 )| c|yk yk+1 | c(dck1 ) = dck , as desired. Thus, by induction, I can conclude that |yn yn+1 | dcn1 for all n. Now, a suitable modication of Problem 5 from the Take-Home part of Midterm #1 (with dcn1 taking the place of (1/2)n ) implies that (yn ) is a Cauchy sequence. 1

(c) Prove that y is a xed point of f (i.e., f (y ) = y ) and that it is unique in this regard. Proof. Since (yn ) y and f is continuous, we know that (f (yn )) f (y ). However, for each n, f (yn ) = yn+1 , so the sequence (f (yn )) is the same as the sequence (yn+1 ), which converges to y . Therefore, we can conclude that f (y ) = y . To see uniqueness, suppose that f (z ) = z . Then, on the one hand, |f (y ) f (z )| c|y z |. On the other hand, |f (y ) f (z )| = |y z | since y and z are xed points. Putting this together with the above inequality, we see that |y z | = |f (y ) f (z )| c|y z |. Since 0 < c < 1, this can only be true if |y z | = 0, meaning that z = y . Since the choice of xed point z was arbitrary, we see that every xed point of f is equal to y , so y is the unique xed point of f . (d) Finally, prove that if x is any arbitrary point in R then the sequence (x, f (x), f (f (x)), . . .) converges to y dened in (b). Proof. The same proof as in (b) implies that (x, f (x), f (f (x)), . . .) converges to some number z . Moreover, the same proof as in the rst part of (c) implies that f (z ) = z . Therefore, the uniqueness part of (c) implies that z = y , so the sequence (x, f (x), f (f (x)), . . .) converges to y , as desired. 3. Exercise 4.3.12. Let C be the Cantor set constructed in Section 3.1. Dene g : [0, 1] R by g (x) = 1 0 if x C if x / C.

(a) Show that g fails to be continuous at any point c C . Proof. Clearly, for any c C , g (c) = 1. On the other hand, the set C is totally disconnected (cf. Exercise 3.4.9). Therefore, for each n = 1, 2, 3, . . ., there exists xn /C such that c 1/n < xn < c + 1/n. Hence, (xn ) is a sequence entirely contained in the complement of C such that (xn ) c. Since g (xn ) = 0 for all n, we see that (g (xn )) does not converge to g (c), so Theorem 4.3.2(iv) implies that g cannot be continuous at c. (b) Show that g is continuous at every point c / C. Proof. Let c C c [0, 1]. Let Therefore, > 0. Since C is a closed set, we know that C c is open. C c [0, 1] = C c (0, 1) is open, since C c and (0, 1) are both open. Hence, there exists > 0 such that V (c) C c [0, 1]. Therefore, if |x c| < , we know that x V (c) C c [0, 1], meaning that |g (x) g (c)| = |0 0| = 0 < . Since the choice of was arbitrary, we conclude that g is continuous at c. In turn, since the choice of c was arbitrary, we see that g is continuous on all of C c [0, 1]. 2

4. Exercise 4.4.9. A function f : A R is called Lipschitz if there exists a bound M > 0 such that f (x) f (y ) M xy for all x, y A. Geometrically speaking, a function f is Lipschitz if there is a uniform bound on the magnitude of the slopes of lines drawn through any two points on the graph of f . (a) Show that if f : A R is Lipschitz, then it is uniformly continuous on A. Proof. Suppose f is uniformly continuous with Lipschitz constant M . Let whenever |x y | < /M , we have that |f (x) f (y )| M |x y | < M M = , > 0. Then

where the rst inequality is equivalent to the Lipschitz condition. Since the choice of was arbitrary, we conclude that f is uniformly continuous. (b) Is the converse statement true? Are all uniformly continuous functions necessarily Lipschitz? Answer. No, not all uniformly continuous functions are Lipschitz. The basic example of a uniformly continuous function which is not Lipschitz is a continuous function on a compact domain which has a vertical tangent line somewhere. For example, consider the function f (x) = 3 x on [1, 1]. We proved in class that f is continuous on all of R, so it is certainly continuous on [1, 1]; since [1, 1] is compact, Theorem 4.4.8 implies that f is uniformly continuous on [1, 1]. However, for xn = 1/n3 and yn = 1/n3 we have that f (xn ) f (yn ) 1/n 1/n 2/n = = = n2 . 3 3 xn yn 1/n 1/n 2/n3 Clearly, there is no M such that n2 M for all n, so the function f cannot be Lipschitz. 5. Exercise 4.4.11 (Topological Characterization of Continuity). Let g be dened on all of R. If A is a subset of R, dene the set g 1 (A) by g 1 (A) = {x R : g (x) A}. Show that g is continuous if and only if g 1 (O) is open whenever O R is an open set. Proof. () Suppose g is a continuous function. Let O R be an open set and let x g 1 (O), which means that g (x) O. Now, since O is open, there exists > 0 such that V (g (x)) O. Moreover, since g is continuous, there exists > 0 such that y V (x) implies that g (y ) V (g (x)) O. This implies that whenever y V (x) we have that g (y ) O, meaning that y g 1 (O). Since the choice of y V (x) was arbitrary, we see that V (x) g 1 (O). Since the choice of x was arbitrary, we see that every element of g 1 (O) has some -neighborhood contained in g 1 (O), so we conclude that g 1 (O) is open. 3

() On the other hand, suppose g 1 (O) is open whenever O R is open. Pick c R and let > 0. Then V (g (c)) is open, so, by hypothesis, g 1 (V (g (c))) is open. Since c g 1 (V (g (c))), opennness implies that there exists > 0 such that V (c) g 1 (V (g (c))). In other words, if y V (c), then g (y ) V (g (c)). Since the choice of > 0 was arbitrary, we see that g is continuous at c. Since the choice of c was arbitrary, we conclude that g is continuous wherever it is dened, namely on all of R.