Exercise 1.9

Suppose μ is a positive measure on X , f : X [ 0 , ] is measurable, X f = c , where 0 < c < and α is a constant. Prove that

lim n X n log [ 1 + ( f n ) α ] = { if 0 < α < 1, c if α = 1, 0 if 1 < α < ∞.

Answers

Proof. Suppose α 1 . Then, by the Taylor series for e x , we have that 1 + x α e αx . Substituting x = f n and applying log , we get the inequality

f n : = n log [ 1 + ( f n ) α ] αf .

Thus, we have that αf f n for all n . Now, by the dominated convergence theorem, we have

lim n X f n = X lim n f n .

We now calculate lim n f n :

lim n f n = lim n n log [ 1 + ( f n ) α ] = lim n log [ 1 + ( f n ) α ] n α n α 1 .

If α > 1 , then the numerator converges to f α using the sequence definition of e x , while the denominator diverges to , hence the limit is zero. On the other hand, if α = 1 , then the denominator is 1 , and so the limit is f α = f . Thus, we have that the integral we are interested in has the desired values for α 1 .

Now suppose α < 1 . By Fatou’s lemma, we have

X liminf n f n liminf n X f n lim n X f n ,

and it suffices to show the integral on the very left diverges. But we have

liminf n f n = liminf n n log [ 1 + ( f n ) α ] = liminf n n 1 α log [ 1 + ( f n ) α ] n α f α liminf n n 1 α =

everywhere f ( x ) 0 , which happens on a set of positive measure since we have X f > 0 . □

User profile picture
2021-12-22 00:00
Comments