Let $m \in \mathbb{N}\setminus \{0,1\}$, $\alpha \in ]0,1[$. Let $\Omega$ be a bounded open subset of $\mathbb{R}^n$ of class $C^{m,\alpha}$.

It is known that if $f \in C^{\frac{m-2+\alpha}{2},m-2+\alpha}([0,T]\times \mathrm{cl}\,\Omega)$, $g \in C^{\frac{m+\alpha}{2};m+\alpha}([0,T]\times\partial\Omega)$, $u_0 \in C^{m,\alpha}(\mathrm{cl}\, \Omega)$ (satisfying some compatibility conditions at $t=0$), then there exists a unique solution $u$ in $C^{\frac{m+\alpha}{2};m+\alpha}([0,T]\times \mathrm{cl}\,\Omega)$ of \begin{cases} \partial_t u -\Delta u = f &\mbox{ in }[0,T]\times \mathrm{cl}\,\Omega,\\ u=g & \mbox{ on } [0,T] \times \partial\Omega,\\ u(0,\cdot) = u_0 & \mbox{ in }\mathrm{cl} \, \Omega. \end{cases}

For the elliptic case holds a very similar result, but in this case we allow $m$ to be also $1$, that is $m \in \mathbb{N}\setminus \{0\}$. If $f \in C^{m-2,\alpha}(\mathrm{cl}\,\Omega)$, $g \in C^{m,\alpha}(\partial\Omega)$, then there exists a unique solution $u$ in $C^{m,\alpha}(\mathrm{cl}\,\Omega)$ of \begin{cases} \Delta u = f &\mbox{ in }\mathrm{cl}\,\Omega,\\ u=g & \mbox{ on } \partial\Omega. \end{cases} In this case, for $m=1$, the space $C^{-1,\alpha}(\mathrm{cl}\,\Omega)$ is the space of distributions which euqals the divergence of an element in $C^{0,\alpha}(\mathrm{cl}\,\Omega,\mathbb{C}^n)$, and the Laplacian is to be intended in the weak sense.

Then my question is the following:

There exists an analog of the case $m=1$ for the heat (parabolic) equation?


2 Answers 2


For a partial answer, see Theorem 6.48 in Second order parabolic differential equations, 1996, by Gary Lieberman.

There is an additional assumption that $f$ belongs to the Morrey space $M^{1,n+1+\alpha}$ defined page 130 of the book. In particular, $L^{\infty}\subset M^{1,n+1+\alpha}\subset L^1$.

I do not know if you can remove this additional assumption or not.

  • 1
    $\begingroup$ Could you include the statement of Theorem 6.48 for those who don't have immediate access to the book? $\endgroup$ Aug 26, 2021 at 12:54
  • $\begingroup$ @MichaelAlbanese: That book is the prime example about how not to write a book if you want it to be read. It contains no enlighting ideas, just boring and heavy technique, and theorems (and notations) always depend on previous chapters and results. I could edit this post to include the stament of theorem 6.48, but you wouldn't understand a thing of it. So much so, that it is impossible to me to decide whether that theorem does answer the question or not... $\endgroup$
    – Alex M.
    Aug 26, 2021 at 13:14
  • $\begingroup$ @AlexM. the theorem does answer the question, you just have to check what means every notation used in the statement (which is not straight forward at all). However, I agree with the fact that there is no way to check the details of the proofs or to rewrite it by yourself. $\endgroup$
    – ZZZ
    Aug 26, 2021 at 13:28
  • $\begingroup$ @MichaelAlbanese, it would take me pages to introduce all the notations used in the statements. Moreover, the functional spaces are not the most usual, so I think that it is better to check directly in the book. $\endgroup$
    – ZZZ
    Aug 26, 2021 at 14:02
  • $\begingroup$ @ZZZ: Fair enough. $\endgroup$ Aug 26, 2021 at 14:21

A partial answer.

There are results for parabolic Schauder theory in $C^{1,\alpha}(\bar \Omega)$. Namely, for the rhs $f\equiv0$, $u_0\in C^{1,\alpha}$, $g\in C^{1,\alpha}$ see Baderko E.A. "Parabolic problems and boundary integral equations", Math. Methods Appl. Sci. 1997, V.20, P. 449-459. The results are for unbounded cylinder $\Omega\times[0,\infty)$ and for parabolic equations with Holder coefficients. There are references here for her earlier works in bounded wrt $t$ cylinders.

As for the rhs $f$, it is shown in M. F. Cherepova, “On some properties of the parabolic potential of bulk masses. I”, Differ. Uravn., 35:12 (1999), 1701–1706; Differ. Equ., 35:12 (1999), 1726–1732, if it satisfies the condition $|f|\le C d^{\alpha-1}$, where $d$ is the parabolic distance to the parabolic boundary, then the volume potential $Vf$ belongs to $C^{1,\alpha}$ in the closure of the domain and therefore solutions of the first BVP with such $f$ (and $u_0$, $g$) belong to the same class.


Your Answer

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

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