Loading [MathJax]/jax/output/HTML-CSS/jax.js

Friday, August 9, 2013

Real Exponents (1.6)

Walter Rudin Principles of Mathematical Analysis, chapter 1, exercise 6:

MathJax TeX Test Page Fix b>1.
(a) Let m,n,p,q be integers, n,q>0 with m/n=p/q. Prove(bm)1/n=(bp)1/qHence it makes sense to define bm/n=(bm)1/n.
(b) Prove br+s=brbs for r,sQ.
(c) For xR, define B(x)={bt | tQ, tx}. When rQ, provebr=sup B(r)Hence it makes sense to definebx=sup B(x)(d) For x,yR, prove bx+y=bxby.

Proof: (a) Recall (z1z2)1/n=z1/n1z1/n2 for natural n. For natural n1,n2 it is also clear that (zn1)n2=zn1n2, and when n1,n2 may be negative the equality holds by case analysis together with the definition zn=(zn)1. Therefore, we derive(bm)1/n=(b1/n)m=(((b1/n)m)q)1/q=((b1/n)mq)1/q=((b1/n)pn)1/q=(((b1/n)n)p)1/q=(bp)1/q (b) The case is clear for integer r,s. As well, we note z1/n1n2=(z1/n1)1/n2 as ((z1/n1)1/n2)n1n2=(((z1/n1)1/n2)n2)n1=z. Otherwise let r=r1/r2 and s=s1/s2. We havebr+s=br1/r2+s1/s2=b(r1s2+r2s1)/(r2s2)=(br1s2+r2s1)1/r2s2=(br1s2br2s1)1/r2s2=(((br1)s2)1/s2)1/r2(((bs1)r2)1/r2)1/s2=brbs (c) Lemma 1: When a=a1/a2>0 is rational we have ba>1. This implies ba>bc for rationals a>c. Proof: Rewrite a2>0 so that a1>0 so by induction ba1>1 and it suffices to prove b1/a2>1 for b>1. Here it suffices to prove for c1 then cn1 for natural n, which is clear by induction cn=ccn1c1.

Now, for tr we show brbtbt(brt1)0, so since bt is nonnegative and by the lemma brt1 is nonnegative we have br is an upper bound of B(r). As well, if z<br then since brB(r) we have z is not an upper bound of B(x) and thus br=sup B(r).

(d) Lemma 2: Let XR be bounded above such that when xX and yx, then yX. We see sup X=sup (X{sup X}). Proof: z=sup X bounds the latter, so assume y<z also bounds. But y doesn't bound X and thus y<x for some xX and we may choose some y<x1<xz so x1X{sup X} yet y<x1 and z is the smallest bound.

Let B(x)={vbt | tQ, tx} so that sup B(x)=sup B(x) and it serves as an equivalent definition for bx. Now if bxB(x) then by the above lemma we may remove it, and as well bt for t<x to define B(x) with an equivalent supremum. Now we must show that when z<sup B(x) then z<bv for some rational v<x. We already see this is the case when for some rational vx, so we must construct rational v<v such that z<bv<bv; this may be done by observing lemma 1 and parts (a),(b), and (g) to show b1/n tends to 1 and setting v=v1/n for 1<b1/n<bv/z.

Now, suppose bxby<bx+y; then bxby<bv for some rational v<x+y by the above discussion. Set rational vy<v1<x and vv1<v2<y so v1+v2>v and bxbybv1bv2=bv1+v2>bv.

Suppose bx+y<bxby. Then (bx)1bx+y<bv2 for some rational v2<y. Choose rational v2<v2<y and set v4=v3v2>0, and then choose xv4<v1<x. We have bv2>(bx)1bx+y(bx)1bv1+v2+v4 implying 1>(bx)1bv1bv4 then bx>bv1+v4. But v1+v4>x and choosing rational x<x<v1+v4 we see bx is an upper bound of bx and bv1+v4>bxbx is a contradiction. 

No comments:

Post a Comment