1

Denote by $$w_{\lambda}(t):=\int_0^t f_{\lambda}(t,\sigma) d\sigma, \quad t \in \mathbb{R},$$ where $f_{\lambda}(t,\sigma)$ is a continuously differentiable function on $\mathbb{R}$ in respect to $t$ and $f_{\lambda}(t,t)=0$, $\lambda \in \mathbb{R}$.

The aim is to prove that $v(t):=\displaystyle\lim_{\lambda \to \infty} w_{\lambda}(t)$ defines a continuously differentiable function on $\mathbb{R}$.

Using Leibniz integral rule, we can prove that $$w_\lambda'(t)=\int_0^t \frac{\partial}{\partial t}f_{\lambda}(t,\sigma) d\sigma.$$ I also proved using some results on the function $f_{\lambda}(t,\sigma)$ that there exists $\omega \geq 0$ such that for all $\lambda>\omega$ we have $$\|w_\lambda'(t)\| \leq \frac{\lambda}{\lambda-\omega} K(T), \quad \forall t \in [-T,T],$$ where $K(T)$ is a positive constant independent of $t$, this means that the limit of $w_\lambda'(t)$ converge uniformly on compacts of $\mathbb{R}$. Can we deduce then that $v(t)$ is a continuously differentiable function on $\mathbb{R}$?

Bruno B
  • 7,655
  • Continuity and differentiability are local properties, so you just have to prove that on each interval your limit is continuously differentiable, which is pretty much already done considering you already have uniform convergence of the derivatives on compact sets, so that you only need pointwise convergence of the $w_\lambda$s at one point, which you can take $0$ for, considering $w_\lambda(0) = 0$. (Just in case, here's a link to a proof of my claim, in case it's not a standard fact: Uniform convergence of derivatives, Tao 14.2.7.) – Bruno B Jul 22 '24 at 13:55

0 Answers0