Define what it means to say that fn→f weakly in L2([0,1]).
Suppose fn∈L2([0,1]) converge weakly to f∈L2([0,1]) and define primitive functions:
Fn(x)=∫0xfn(t)dtandF(x)=∫0xf(t)dt.
Show that Fn,F∈C([0,1]) and that Fn→F uniformly on [0,1].
Solution.
fn→f weakly in L2([0,1]) if for every g∈L2([0,1]), we have
∫fngdtn→∞∫fgdt.
Notice that by weak convergence, supn∣⟨fn,g⟩∣<∞ for any g∈L2([0,1]). L2([0,1]) is its own dual, we may view {fn}n⊆(L2([0,1]))∗ and so by the uniform boundedness principle,
Notice that the upper bound is uniform over n≥1, so {Fn}n is actually equicontinuous. Similarly,
∣Fn(x)∣≤∫01∣fn(t)∣dt≤∥fn∥L2≤M,
so {Fn}n is uniformly bounded. By Arzelà-Ascoli, {Fn}n admits a uniformly convergent subsequence. Moreover, since χ[0,x]∈L2([0,1]), weak convergence also implies
Thus, any convergent subsequence of {Fn}n must converge to F. This implies that Fn→F uniformly itself. Otherwise, there exists ε>0 such that there exists xn1∈[0,1] with ∣Fn1(xn1)−F(xn1)∣≥ε. By induction, we get a subsequence nk such that ∣Fnk(xnk)−F(xnk)∣≥ε for all k≥1, so no subsequence of {Fnk}k may converge uniformly to F. But by Arzelà-Ascoli, {Fnk}k must admit a uniformly convergent subsequence to F, a contradiction. Hence, Fn→F uniformly to begin with, which completes the proof.