Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
1K views

Abbott - Understanding Analysis Problem Set

The document contains solutions to various problems in mathematical analysis: 1) It proves that the composition of two continuous functions is continuous using both epsilon-delta and sequential definitions of continuity. 2) It proves that Dirichlet's function, which is 1/x for rational x and 0 for irrational x, is nowhere continuous. 3) It shows that the preimage of zero under a continuous function is a closed set. 4) It gives an example of a continuous and positive function on a closed interval that is bounded above but not below.

Uploaded by

Feynman Liang
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
1K views

Abbott - Understanding Analysis Problem Set

The document contains solutions to various problems in mathematical analysis: 1) It proves that the composition of two continuous functions is continuous using both epsilon-delta and sequential definitions of continuity. 2) It proves that Dirichlet's function, which is 1/x for rational x and 0 for irrational x, is nowhere continuous. 3) It shows that the preimage of zero under a continuous function is a closed set. 4) It gives an example of a continuous and positive function on a closed interval that is bounded above but not below.

Uploaded by

Feynman Liang
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Feynman Liang MATH 355 Assignment #7 11/4/2013

Abbott: 4.3.2, 4.3.6(a), 4.3.7, 4.4.4, 4.4.6, 4.4.8, 4.4.13(a)


4.3.2 Given f : A R, g : B R, f ( A) B s.t. g f ( x ) is well dened on A. f cts at c A, g cts at f (c) B = g f cts at c. (a) Supply a proof using - characterization of continuity. (b) Give another proof using sequential characterization of continuity (Theorem 4.3.2 (iv)). (a) If c is an isolated point, then N (c) : N (c) A = {c}. Then since c is the only element in N (c), x N (c), g f ( x ) = g f (c) N ( g f (c)) for any > 0, showing g f cts (Thm 4.3.2). Thus, assume c is a limit point. Since f cts at c, c A. Let > 0 be arbitrary. Since g cts at f (c), 0 > 0 s.t. | f ( x ) f (c)| < 0 = | g f ( x ) g f (c)| < . Since f cts at c, 0 > 0 s.t. | x c| < 0 = | f ( x ) f (c)| < 0 . Taken together:

| x c| < 0 = | f ( x ) f (c)| < 0 = | g f ( x ) g f (c)| <


So x N0 (c) = g f ( x ) N ( g f (c)). Since arbitrary, by Theorem 4.3.2 g f cts. (b) If c is an isolated point, then for any sequence ( xn ) A converging to c, N0 N : n N0 , xn = c. This implies n N0 , g f ( xn ) = g f (c) and therefore ( g f ( xn )) g f (c), proving g f cts (Thm 4.3.2). Thus, assume c is a limit point. Let ( xn ) be an arbitrary sequence s.t. ( xn ) c, c A, f cts at c. Since f cts at c , ( xn ) c = ( f ( xn )) f (c). Since g cts at f (c) and ( f ( xn )) f ( A) B, ( f ( xn )) f (c) = ( g f ( xn )) g f ( c ). Taken together:

( xn ) c = ( f ( xn )) f (c) = ( g f ( xn )) g f (c)
So ( xn ) c = ( g f ( xn )) g f (c). Since ( xn ) and c arbitrary, by Theorem 4.3.2 g f cts. 4.3.6(a) Prove that Dirichlets function: g( x ) = is nowhere continuous. 1 1 0 if x Q if x Q

Let x R arbitrary. Case I: x Q Since Q is dense in R, > 0, N ( x ) Q = . Thus, we can produce a sequence which converges to x by letting n N be arbitrary and choosing any arbitrary xn N1/n ( x ), which is non-empty. Notice that ( xn ) x and g( x ) = 0. However, n N, xn Q so g( xn ) 1. Thus by Corrolary 4.3.3, g is not cts at x. Since x R arbitrary, g is nowhere cts. Case II: x Q Since R \ Q is dense in R, > 0, N ( x ) (R \ Q) = . Thus, we can produce a sequence which converges to x by letting n N be arbitrary and choosing any arbitrary xn N1/n ( x ). Notice that ( xn ) x and g( x ) = 1. However, n N, xn (R \ Q) so g( xn ) 0. Thus by Corrolary 4.3.3, g is not cts at x. Since x R arbitrary, g is nowhere cts. 4.3.7 Assume h : R R cts and let K = { x : h( x ) = 0}. Show that K is closed. If K = then it is closed so assume that K is non-empty. Let c be an arbitrary limit point of K, which implies ( xn ) (K \ {c}) : ( xn ) c. Then since h cts, (h( xn )) h(c). Furthermore, notice that ( xn ) K so h( xn ) = 0, n N. Thus, (h( xn )) = (0, 0, 0, . . .) and (h( xn )) 0, implying h(c) = 0. Since h(c) = 0, then c K. Since c arbitrary, K contains all its limit points and is closed. 4.4.4 Prove f cts on [ a, b] , x [ a, b] : f ( x ) > 0 = 1 bdd on [ a, b] f

[ a, b] is closed and bounded, so by HBT it is compact. Since f is cts on [ a, b], EVT (Theorem 4.4.3) implies that that: x0 [ a, b ] : x [ a, b ], f ( x0 ) f ( x )
Since f ( x ) > 0 for x [ a, b], f ( x0 ) > 0 as well. Inverting the inequality yields x [ a, b], f (1 f (1x) , showing 1 f bdd. x )
0

4.4.6 Give an example or show impossibility: (a) f : (0, 1) R cts and ( xn ) Cauchy s.t. f ( xn ) is not Cauchy. (b) f : [0, 1] R cts and ( xn ) Cauchy s.t. f ( xn ) is not Cauchy. (c) f : [0, ) R cts and ( xn ) Cauchy s.t. f ( xn ) is not Cauchy. (d) f : (0, 1) R cts, bdd that attains a maximum value but not a minimum on (0, 1). 2

1 (a) Let f : x 1 x , xn = n . Then f is cts on (0, 1) and ( xn ) 0 ( xn ) Cauchy. However, f ( xn ) = n so ( f ( xn )) diverges ( f ( xn )) not Cauchy.

(b) Impossible. ( xn ) Cauchy ( xn ) x. ( xn ) [0, 1] and [0, 1] closed, so x [0, 1]. Since f cts in [0, 1] and x [0, 1], Theorem 4.3.2 implies ( f ( xn )) f ( x ). Since ( f ( xn )) converges, it is Cauchy. Thus, ( f ( xn )) is Cauchy for all ( xn ) Cauchy. (c) Impossible. Same proof as (b), only changing domain from [0, 1] to [0, ) which is still closed and hence if ( xn ) x is a Cauchy sequence in [0, ) than x [0, ). The rest of the proof is identical. (d) Let f : x x (1 x ). Then the maximum occurs when 0 = 1 2x = x = 1 1 2 , yielding a maximum value for f of f (1/2) = 4 . However, any proposed minimum x0 m0 = f ( x0 ) is not minimal since if m0 0.5 then f ( x0 + 1 2 ) is smaller and similarily if x0 m0 < 0.5 then f ( x0 1 2 ) is smaller. 4.4.8 (a) Assume f : [0, ) R cts on its domain. Show

b > 0 : f uniformly cts on [b, ) = f uniformly cts on [0, )


(b) Prove f : x

x is uniformly cts on [0, ).

(a) Let > 0 be arbitrary. [0, b] is compact (HBT) and so by Theorem 4.4.8 f is uniformly cts on [0, b]. f uniformly cts on [0, b] implies:

1 > 0 : c [0, b], x N1 (c) = f ( x ) N ( f (c))


Similarly, f uniformly cts on [b, ) implies:

2 > 0 : c [0, ), x N2 (c) = f ( x ) N ( f (c))


Choosing = max{1 , 2 }, we get:

c [0, ) = [0, b] [b, ), x N (c) = f ( x ) N ( f (c))


Since

> 0 arbitrary, f is uniformly cts on [0, ). > 0 and x, c 1 be arbitrary. Choose = . Then x : | x c| < x+ y c = ( x y) x+ y =
xy x+ y

(b) Consider [1, ). Let :

| f ( x ) f (c)| =

< | x y| ==
3

Since > 0 and x, c [1, ) arbitrary, f is uniformly cts on [1, ). By part (a), f is uniformly cts on [0, ).

4.4.13(a) Show that if f : A R is uniformly cts and ( an ) A Cauchy, then f ( an ) is Cauchy. Let f : A R be uniformly cts on A and ( xn ) A an arbitrary Cauchy sequence. Let > 0 be arbitrary. f uniform cts means > 0 : x, y A, | x y| < = | f ( x ) f (y)| < . an Cauchy means N0 N : m > n N0 , | am an | < . Letting x = am , y = an in the uniform cts statement above, since ( an ) A we have:

m > n N0 , | f ( am ) f ( am )| <
Since

> 0 arbitrary, ( f ( an )) is Cauchy.

You might also like