Some Convergence Concepts
[12/50%]This result concerns uniqueness.
If -a.e. and -a.e., then -a.e..
If in -measure and in -measure, then -a. e..
Suppose that and .
and .
If , then .
If , then .
and .
If and , there is with .
If , then .
If , then .
If , then does not necessarily converge to in -measure.
It is not necessarily the case that , even if and 0 for every and . However, if , the result follows.
Suppose that and .
and .
If , then .
If , then .
and .
.
If -a. e. for each , then .
Suppose for each that -a. e..
If , then .
If , then .
Prove the following statements connecting convergence in -measure with convergence -a. e.
iff for every subsequence .
iff each subsequence of has a sub-subsequence that converges to -a. e..
Suppose that is continuous.
If , then .
If , then .
Let , let , and let denote the restriction of Lebesgue measure to . Let , , , , , , , , etc. On , define for each .
in -measure.
For each , , hence fails to exist for any . Therefore, the sequence converges to no function with domain . In particular, fails to converge -a.e. to , so that convergence in -measure does not imply convergence -a.e., even if .
Consider the measure space , and let for each and .
-a.e.
fails to converge in -measure to , hence the hypothesis in Claim 3 is not redundant.
If and , then cannot converge uniformly to on . This shows that the hypothesis is not redundant in Claim 4.
Claim 3 and Claim 4 referenced here are from earlier in this section: Claim 3 states that if is a measure space with , and each of is -measurable with domain and range contained in , such that -a.e. on , then in -measure on . Claim 4 (Egorov's Theorem) states that under the same hypotheses (with ), for every there exists such that and uniformly on .
If for each , then -a.e.
We have assumed throughout Section 5.6 that the functions have domain and are finite-valued. We may generalize the previous definitions and discussion to the case where each of is defined -a.e.on and each of is finite -a.e.on . Let denote the set of such that (i) is defined and finite, and (ii) is defined and finite for each . Then
and is a countable union of -null sets, so . We may therefore decompose into the two disjoint -sets and , and think of as denoting -measurable functions with domain and range contained in , where denote the restrictions of (respectively) from their respective domains to . We say that -a.e.on iff , i.e., iff -a.e.on ; and we say that in -measure on iff in -measure on , i.e., iff for every and we have .
State and prove Claim 2 in this more general setting. Mimic the spirit of the example given earlier for Claim 1's own generalization: Claim 1 states that if denotes a measure space and , and if are -measurable functions defined and finite -a.e.on such that in -measure, then there exists a subsequence of such that -a.e.on ; this is proved by applying the original (everywhere-defined, finite-valued) version of Claim 1 to on to get a subsequence with -a.e.on , which by definition means -a.e.on .
Claim 2 (the result to be generalized here): Let denote a measure space with . Pick nonempty . Suppose that (i) each of is -measurable with domain and range contained in , and (ii) -a.e.on . Then, given any and , we may find a set and such that and for every and .
State and prove Claim 3, from earlier in Section 5.6, in the more general -a.e.-on-/-measure-on- setting introduced in problem 5.6.10 (where each of is only assumed defined and finite -a.e.on , rather than everywhere on ).
Claim 3 (the result to be generalized here): Let denote a measure space with , and let . Suppose that each of is -measurable with domain and range contained in such that -a.e.on . Then in -measure on .
State and prove Claim 4 (Egorov's Theorem), from earlier in Section 5.6, in the more general -a.e.-on-/-measure-on- setting introduced in problem 5.6.10 (where each of is only assumed defined and finite -a.e.on , rather than everywhere on ).
Claim 4 (Egorov's Theorem, the result to be generalized here): Let denote a measure space with , and let be nonempty. Let each of be -measurable functions with domain and range contained in such that -a.e.on . Then for every there exists such that and uniformly on . (Roughly speaking, this says that if -a.e., then uniformly except on a set of arbitrarily small -measure — that is, "uniformly almost everywhere.")