Homepage Solution manuals Terence Tao An Introduction to Measure Theory Exercise 1.5.7 (Pointwise convergence in double index)

Exercise 1.5.7 (Pointwise convergence in double index)

Let (X,,μ) be a measure space, let (f )n be a sequence of measurable functions from X to converging pointwise almost everywhere to a measureble limit f : X as n , and for each n let (fn,m) n,m be a sequence of measurable functions from X to converging pointwise almost everywhere to fn as m .

(i)
If μ(X) < , show that there exists a sequence m1,m2, such that fn,mn converges pointwise almost everywhere to f as n .
(ii)
Show that the same claim is true if X is σ-finite.

Answers

(i)
By Theorem 1.5.9 (Egorov’s theorem for finite measure spaces) the conditions of (f )n f and (fn,m) n,m pointwise almost everywhere are equivalent to convergence in almost uniform mode. In other words, except for some exceptional set E0 B with m(E0) < δ2 for an arbitrary δ > 0 we have 𝜖 > 0N n Nx XE0 : d(fn(x),f(x)) 𝜖 2

Similarly, for each n we can find an exceptional set En B of measure δ2 2n such that

𝜖 > 0mn m mnx XEn : d(fn,m(x),fn(x)) 𝜖 2

Thus consider a subsequence (fn,mn) n constructed from (fn,m) n,m by axiom of choice using above logic. On the exceptional set E := n=0En with μ(E) δ2 + 12 i=1nδ2n = δ, we then have that

For N𝜖 n N𝜖x XE : d(fnm(x),f(x)) d(fnm(x),fn(x))+d(fn(x),f(x)) 𝜖

In other words, (fn,mn) n converges to f almost uniformly, which, by Egorov’s theorem, is equivalent to almost everywhere convergence pointwise.

(ii)
This follows by applying the above proof countable amount of times to the partition X = n=1Xn of σ-measurable sets Xn with μ(Xn) < using almost uniform convergence and the 𝜖2n-trick.
User profile picture
2020-12-09 00:00
Comments
  • Your sequence $m_n$ is dependent on $\varepsilon$.
    isn2025-05-29