User:IssaRice/Taking inf and sup separately

From Machinelearning

This page describes a trick that is sometimes helpful in analysis.

Satement

Let A and B be bounded subsets of the real line. Suppose that for every aA and bB we have ab. Then inf(A)sup(B).

Actually, do A and B have to be bounded? I think they can even be empty!

Proof

Let aA and bB be arbitrary. We have by hypothesis ab. Since b is arbitrary, we have that a is an upper bound of the set B, so taking the superemum over b we have asup(B) (remember, sup(B) is the least upper bound, whereas a is just another upper bound). Since a was arbitrary, we see that sup(B) is a lower bound of the set A. Taking the infimum over a, we have inf(A)sup(B), as required.

Applications

liminf vs limsup

Let (an)n=m be a sequence of real numbers. Let L:=lim infnan and let L+:=lim supnan. Then we have LL+.

Consider the sequences (aN)N=m and (aN+)N=m defined by aN:=inf(an)n=N and aN+:=sup(an)n=N.

Now consider the sets A:={aN+:Nm} and B:={aN:Nm}. If we can show that aj+ak for arbitrary j,km, then we can apply the trick to these sets to conclude that L+=inf(aN+)N=minf(A)sup(B)=sup(aN)N=m=L.

Lower and upper Riemann integral