Scribd
Upload
21 views

Project 3.1

Kaitlyn Monteer proves Theorems 3.8 and 3.11 without using compactness. For Theorem 3.8, which states that a continuous function on a closed and bounded domain is uniformly continuous, Kaitlyn assumes it is not uniformly continuous and derives a contradiction. For Theorem 3.11, which states that a continuous function on a closed and bounded domain achieves its minimum and maximum, Kaitlyn shows the function's range is closed and bounded, so it must achieve its infimum and supremum.

Uploaded by

Kaitlyn M
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
21 views

Project 3.1

Kaitlyn Monteer proves Theorems 3.8 and 3.11 without using compactness. For Theorem 3.8, which states that a continuous function on a closed and bounded domain is uniformly continuous, Kaitlyn assumes it is not uniformly continuous and derives a contradiction. For Theorem 3.11, which states that a continuous function on a closed and bounded domain achieves its minimum and maximum, Kaitlyn shows the function's range is closed and bounded, so it must achieve its infimum and supremum.

Uploaded by

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

Project 3.

Kaitlyn Monteer

In this project, I want to prove versions of Theorem 3.8 and 3.11 without using the
notion of compactness, meaning I will not use any theorems which rely on compactness as
justication in my proofs. These theorems are as follows:
THEOREM 3.8: Suppose f : A R is continuous with A closed and bounded. Then f
is uniformly continuous.
Let f : A R be continuous, with A closed and bounded. Suppose, by way of
contradiction, that f is not uniformly continuous on A. Then there exists some > 0 such
that for all > 0, there exist x, y with |x y| < and |f (x) f (y)| . We will consider
sequences satisfying these conditions.
Proof.

1
So, let > 0 be given and let {n }
n=1 be given by n = n . Now, by our assumption we can

nd {an }
n=1 , {bn }n=1 such that an , bn A and |an bn | < n for all n, and
|f (an ) f (bn )| .

Now, by our denition of the sequences {an }


n=1 and {bn }n=1 , an , bn A for all n. Hence,

since A is bounded, {an }


n=1 and {bn }n=1 are bounded sequences. By exercise 1.36 (pg. 57),
every bounded sequence has a convergent subsequence. Hence there exists a sequence of

positive integers {nk }


k=1 such that {ank }k=1 and {bnk }k=1 both converge. Call the limits of
these sequences a and b, respectively.

Now, let = {ank |k J}. Clearly, if a , then a A, since A.


If, however, a / , then ank 6= a for all k. By the denition of convergence, for all > 0,
there exists a K J such that for all k K ,
|ank a| < .

So,
a < ank < a +

So, since ank 6= a for all k, this means that for all > 0, the interval (a , a + ) contains
an element of A that is dierent from a, so by the lemma on page 39, a is an accumulation
point of A. So, since A is closed, a A.
So, in either case, we see that a A
By a similar argument, letting = {bnk |k J}, we nd that b A.

Now, we dened {ank }


k=1 and {bnk }k=1 such that

|ank bnk | <

1
nk

k1 , so |ank bnk | < k1 .


[Note: this property is inherited from our denition the sequences {an }
n=1 and {bn }n=1 .]

So, suppose a 6= b. Then |a b| > 0, so there exists some N J such that


|a b| >

1
N

But, by the denition of convergence, for each > 0, there exist K1 , K2 J such that
|ank a| < 2 for all k K1 , and
|bnk b| < 2 for all k K2 .

Now, let K = max{N, K1 , K2 }, and suppose k K . Then


|a b| = |(ank a) + (bnk b) (ank bnk )|
|ank a| + |bnk b| + |ank bnk |
1
< + +
2 2 k
1
<+ ,
N

since k N, K1 , K2 . Since this is true for all > 0, it follows that


|a b|

1
.
N

This is a contradiction to our assumption that |a b| > N1 , hence a = b.

Observe that since f is continuous and {ank }


k=1 a and {bnk }k=1 b, it follows that

{f (ank )}k=1 f (a) and {f (bnk )}k=1 f (b), by Theorem 3.1. But since a = b,
f (a) = f (b), so {f (bnk )}
k=1 f (a). By the denition of convergence, for each > 0, there
exist K1 , K2 J such that

|f (ank ) f (a)| < 2 for all k K1 , and


|f (bnk ) f (a)| < 2 for all k K2 .

Let K = max{K1 , K2 } and let k K . Then


|f (ank ) f (bnk )| = |(f (ank ) f (a)) (f (bnk ) f (a))|
|f (ank ) f (a)| + |f (bnk ) f (a)|

< +
2 2
= ,

since k K1 , K2 . This is a contradiction to our earlier assumption that


2

|f (ank ) f (bnk )| .

Hence, we must reject the earlier assumption, meaning that f must be uniformly
continuous on A.


Next, I want to prove Theorem 3.11.


THEOREM 3.11: If f : A R is continuous and A is closed and bounded, then there
are x1 , x2 A such that f (x1 ) f (x) f (x2 ) for all x A.
Let f : A R be a continuous function with A closed and bounded. Then by
Theorem 3.9, f (A) is bounded. Next, we want to prove that f (A) is closed.
Proof.

Let a be an accumulation point of f (A). Then there exists some sequence {an }
n=1 such

that {an }n=1 a, an f (a), and an 6= a for all n, by Theorem 1.17. Since an f (a) for all
n, there also exists a sequence {bn }
n=1 such that bn A for all n and f (bn ) = an for all n.
By exercise 1.36, every bounded sequence has a convergent subsequence. Note that since A
is bounded and bn A for all n, {bn }
n=1 is also bounded. Hence there exists a convergent

subsequence {bnk }k=1 . Say {bnk }k=1 b. By a similar argument to the one in the above
proof of Theorem 3.8, we know that b A.

Now consider the sequence {f (bnk )}


k=1 = {ank }k=1 , a subsequence of {ank }k=1 . Note that
since all subsequences of a convergent sequence converge to the same limit as the parent

sequence, {f (bnk )}
k=1 a. But, since f is continuous, {f (bnk )}k=1 f (b). Hence,
a = f (b). Since f (b) f (A), a f (A). Hence, f contains all of its accumulation points, so
f (A) is closed.

Since f (A) is bounded, it has both an inmum and a supremum. Let s1 , s2 R such that
s1 = inf [f (A)], s2 = sup[f (A)]. By exercise 1.22, either s1 f (A) or s1 is an accumulation
point of f (A). Either way, since f (A) is closed, it contains all of its accumulation points, so
s1 f (A). Similarly, since either s1 f (A) or s1 is an accumulation point of f (A), then
s2 f (A). So, since s1 , s2 f (A), there exists x1 , x2 A such that s1 = f (x1 ), x2 = f (x2 ).
Finally, by the denitions of the supremum and inmum, if x A, f (x) f (A), then
f (x1 ) f (x) f (x2 ).


You might also like