27.08.2014 Views

Real and Complex Analysis (Rudin)

Real and Complex Analysis (Rudin)

Real and Complex Analysis (Rudin)

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

272 REAL AND COMPLEX ANALYSIS<br />

m = max {I z - IX I: Z E K}.<br />

I p(z) - f(z) I < 11m for all z E K. Then<br />

Suppose P is a polynomial, such that<br />

I (z - IX)P(Z) - 11 < 1 (z E K). (1)<br />

In particular, (1) holds if z is in the boundary of V; since the closure of V is<br />

compact, the maximum modulus theorem shows that (1) holds for every<br />

z E V; taking z = IX, we obtain 1 < 1. Hence the uniform approximation is<br />

not possible.<br />

The same argument shows that none of the IXi can be dispensed with in<br />

Theorem 13.6.<br />

We now apply the preceding approximation theorems to approximation<br />

in open sets. Let us emphasize that K was not assumed to be connected in<br />

Theorems 13.6 <strong>and</strong> 13.7 <strong>and</strong> that n will not be assumed to be connected<br />

in the theorem which follows.<br />

13.9 Theorem Let n be an open set in the plane, let A be a set which has one<br />

point in each component of S2 - n, <strong>and</strong> assume f E H(n). Then there is a<br />

sequence {Rn} of rational functions, with poles only in A, such that Rn - f uniformly<br />

on compact subsets ofn.<br />

In the special case in which S2 - n is connected, we may take A = {oo}<br />

<strong>and</strong> thus obtain polynomials P n such that P n - f uniformly on compact subsets<br />

ofn.<br />

Observe that S2 - n may have uncountably many components; for instance,<br />

we may have S2 - n = {oo} u C, where C is a Cantor set.<br />

PROOF Choose a sequence of compact sets Kn in n, with the properties specified<br />

in Theorem 13.3. Fix n, for the moment. Since each component of<br />

S2 - Kn contains a component of S2 - n, each component of S2 - Kn contains<br />

a point of A, so Theorem 13.6 gives us a rational function Rn with poles<br />

in A such that<br />

1<br />

I Rn(z) - f(z) I < - n<br />

(z E KJ. (1)<br />

If now K is any compact set in n, there exists an N such that K c Kn for<br />

all n ~ N. It follows from (1) that<br />

which completes the proof.<br />

1<br />

I Rn(z) - f(z) I < - n<br />

(z E K, n ~ N), (2)<br />

IIII

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!