Let $\Omega=\{(x,y) \in \mathbb{R}^2: x^2+y^2<1\}$, and let $u\in C^1(\bar{\Omega})$ satisfy

$$\alpha(x,y)u_x+\beta(x,y)u_y =-u \hspace{0.2in} \forall (x,y)\in\bar{\Omega}, $$

where $\alpha$ and $\beta$ are continuous functions on $\bar{\Omega}$ such that

$$\alpha(x,y)x+\beta(x,y)y >0 \hspace{0.2in} \forall (x,y)\in\bar{\Omega}, $$

Prove that $u\equiv 0$.


Let $X_0$ be a maximum of $u$.

If $X_0$ is interior, then $ u(X_0)=-\alpha(X_0) u_x(X_0) -\beta(X_0) u_y(X_0) =0 $ since $\nabla u(X_0)=0$, and therefore $u\le 0$.

If $X_0$ is in the boundary, then since $v=(\alpha(X_0),\beta(X_0))$ is an exterior vector to the ball at $X_0$ (i.e. $v\cdot X_0>0$), we have $\nabla u(X_0)\cdot v \ge 0$. That is $u(X_0)\le 0$ and thus again $u\le0$.

Similarly you can see that $u\ge 0$.

  • $\begingroup$ Why $\nabla u(X_0)\cdot v =\alpha(x,y)u_x+\beta(x,y)u_y \geq 0$? $\endgroup$ – Worawit Tepsan Oct 9 '13 at 17:49
  • $\begingroup$ $\nabla u(X_0)\cdot v=\lim_{t\rightarrow 0^-}\frac{u(X_0+tv)-u(X_0)}{t}\ge 0$. $\endgroup$ – Pocho la pantera Oct 9 '13 at 18:23
  • $\begingroup$ I get it now. Thank you so much. Next step, we also do the same with minimum. We can conclude from the minimum that $u\geq0$. Therefore, $u\equiv 0$ $\endgroup$ – Worawit Tepsan Oct 9 '13 at 19:16
  • $\begingroup$ You are welcome. $\endgroup$ – Pocho la pantera Oct 9 '13 at 19:37
  • $\begingroup$ I have a question. Is $t\to 0^+$? (Because $X_0+tv$ shoud goes inside the ball.) $\endgroup$ – Worawit Tepsan Oct 9 '13 at 19:58

Suppose not. Then either $\max_{\overline{\Omega}} u>0$ or $\min_{\overline{\Omega}} u<0$. Assume it's the first one. The PDE implies that $u$ cannot have an interior critical point with nonzero value. Therefore, the maximum is at some boundary point $(x,y)$. But then the derivative along the (inward-pointing) vector $-(\alpha,\beta)$ is positive at that point, which is impossible.

  • $\begingroup$ If $u\equiv -1$, then $u \not \equiv 0$ and $\max u<0$. $\endgroup$ – Pocho la pantera Oct 9 '13 at 14:06
  • $\begingroup$ Thank you so much for help me solve my problem. $\endgroup$ – Worawit Tepsan Oct 9 '13 at 19:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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