Skip to main content
1 of 4
Isaac
  • 3.5k
  • 16
  • 22

If $\frac{\partial f}{\partial t}(x,t)$ exists a.e and $\frac{\partial^2 f}{\partial t \partial x }$ is continuous, can we improve a.e existence?

The question is as in the title.

Let $f(t,x) : [0,1]^2 \to \mathbb{R}$ be a function which is smooth w.r.t $x$ for each fixed of $t$ and whose derivatives w.r.t $x$ are all absolutely continuous w.r.t $t$.

Next, assume that $\frac{\partial^2 f}{\partial t \partial x}(t,x)$ has a version continuously defined everywhere on $[0,1]^2$.

Lastly, suppose that \begin{equation} \frac{\partial f}{\partial t}(t,x)=\text{some linear combinations of derivatives of }f(x,t) \text{ w.r.t } x \end{equation} holds almost everywhere on $[0,1]^2$.

Then, I wonder if we can conclude that $\frac{\partial f}{\partial t}(t,x)$ can be defined everywhere continuously. Moreover, can we even proceed further and conclude that $f(x,t)$ is smooth "jointly" in $x$ and $t$, by iteration argument?

This seems like a subtle issue to me.. Could anyone please help?

Isaac
  • 3.5k
  • 16
  • 22