0
$\begingroup$

Let the sequence $u_n\in L^2(0,\infty)$ weakly converges to $u\in L^2(0,\infty)$. What can we say about the corresponding Laplace transforms $U_n(s)$ and $U(s)$?

  1. $U_n(s)$ converges point-wise to $U(s)$ for almost all $s>0$.

  2. The convergence in (1) but also uniform.

$\endgroup$

1 Answer 1

1
$\begingroup$

The notation in the question for Laplace transform should be improved for clarity.

  1. The Laplace transforms $L(u_n)(s)=\int_0^\infty e^{-sx}u_n(x)\, dx$ converge to $L(u)(s)$ pointwise for each $s>0$ by the definition of weak convergence.

  2. The convergence need not be uniform. E.g. take $u(x)=x/(1+x^2) \in L^2(0,\infty)$ and $u_n(x)=u(x)$ for $x \in (0,n)$, with $u_n(x)=0$ for $x \ge n$. Then $u_n \to u $ in $L^2(0,\infty)$, yet $\sup_{s>0} [L(u)-L(u_n)](s)=\int_n^\infty u(x) \, dx=\infty$.

$\endgroup$
3
  • $\begingroup$ Is the converse also true? That is, if $L(u_n)(s)$ converges pointwise to $L(u)(s)$, then $u_n(t)$ converges weakly to $u(t)$? $\endgroup$
    – Saj_Eda
    Nov 14, 2019 at 16:24
  • $\begingroup$ If the above answered your question, you can indicate that by accepting the answer. meta.mathoverflow.net/questions/3724/how-to-accept-answer $\endgroup$ Nov 14, 2019 at 21:38
  • $\begingroup$ The comments are not really designed for asking additional questions. Without further assumptions the converse does not hold, consider $u_n(x)=n/x$ for $x>n$ and $u_n(x)=0$ elsewhere. Then $L(u_n)$ tends to $L(0)=0$ pointwise for $s>0$, but $\int_0^\infty u_n u_1\, dx=1$ for all $n$. $\endgroup$ Nov 14, 2019 at 21:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.