MATH 321 - HOMEWORK #8 Due Friday, March 18. We proved the Stone-Weierstrass theorem for intervals [a, b] only, but you may assume that the same theorem holds for any compact set K as stated in the textbook. The textbook does not require the algebra to be unital, but it adds the extra condition that A should not vanish at any point. P ROBLEM 1. (a) Let f : [a, b] → R be a strictly increasing continuous function. Let A = R[f] = { n X ak fk |n ≥ 0, ak ∈ R}. k=0 Prove that the uniform closure of A is C[a, b]. (By f0 we mean the constant function 1.) (b) f, g : [a, b] → R be two continuous functions, and let A = R[f, g] = { n X ak,l fk gl |n ≥ 0, ak,l ∈ R}. k,l=0 Prove that {f, g} separates points if and only if A separates points. In that case show that the uniform closure of A is C[a, b]. P ROBLEM 2. Let K = [a, b] × [c, d] ⊂ R2 . (a) Prove that every continuous function f : K → R is a uniform limit of functions of the form n X hi (x)gi (y), i=1 where hi and gi are continuous functions on [a, b] and [c, d], respectively. The number of terms n is arbitrarily large, but finite. (b) Prove the 2-variable Weierstrass approximation theorem: the uniform closure of R[x, y] ⊂ C(K) equals C(K). P ROBLEM 3. Let A ⊂ C[a, b], where a 6= b, satisfy the properties (1)If f, g ∈ A and c ∈ R, then cf, c + f, max(f, g) ∈ A (2)A separates points. Prove that then the uniform closure of A is C[a, b]. P ROBLEM 4. A special case of Tietze’s extension theorem states that if K ⊂ Rn is a compact set, then every continuous function f : K → R is the restriction of a continuous function F : Rn → R. Prove this theorem as follows. 1 (a) Prove that every continuous function f can be uniformly approximated by restrictions of continuous functions F. To do this, let A be the set of all functions K → R that are restrictions of continuous functions Rn → R and apply Stone-Weierstrass theorem to this A. (b) If f is bounded by M, prove that the approximations F as in the previous part can be chosen also bounded by M. You may assume that if F, G are continuous, then min(F, G) and max(F, G) are also continuous. (c) To prove Tietze’s theorem, start with f and choose the first approximation F1 , such that |f(x) − F1 |C (x)| < for every x ∈ K. Now find an approximation F2 of the error f − F1 |C , such that F2 is bounded by and F2 is 2 -close to f − F1 |C . Continue like this. P In the end, the sum Fi gives the correct function that restricts to f. Tietze’s extension theorem is true more generally for any closed set K in any metric space.