Is the identity
$$\delta(x-a)=\lim_{\epsilon \to 0}\frac{1}{\sqrt{2\pi \epsilon}}\exp{[-\frac{1}{2\epsilon}(x-a)^2]}$$
correct? If yes, what is the proof?
Is the identity
$$\delta(x-a)=\lim_{\epsilon \to 0}\frac{1}{\sqrt{2\pi \epsilon}}\exp{[-\frac{1}{2\epsilon}(x-a)^2]}$$
correct? If yes, what is the proof?
Copyright © 2021 JogjaFile Inc.
Pretend for the moment that $a = 0$. Show that these functions give a family of approximate identities, that is, a family $(f_n)$ of functions that
(In fact, any nonnegative $L^1$ function rescaled in the way you describe will give such a family.)
A family of approximate identities will act as a delta function when integrated any sufficiently nice function, such as the smooth, compactly supported functions. That is, they will satisfy $$\lim \limits_{n \rightarrow \infty} \int f_n(x) g(x) \, dx = g(0)$$ for such $g$.