Suppose that $\Omega\in\mathbb{R}^n$ is bound and open with smooth boundary $\partial\Omega$. Consider the following initial-boundary value problem for a nonlinear heat equation: $$ \begin{cases} %\vspace{3mm} u_t=\Delta u+u(1-u), \ \ &x\in\Omega,\ \ t>0,\\ %\vspace{3mm} u(x,0)=u_0(x),\ \ &x\in\Omega,\\ u=0, \ \ &x\in\partial\Omega,\ \ t>0, \end{cases} $$ where $0\le u_0\le1$. Show that $0\le u(x,t)\le1$, $x\in\Omega$, $t>0$.

We prove $u(x,t)\le1$ by contradiction. Assume for some $(x_1,t_1)$ with $x_1\in\Omega$ and $T>t_1>0$, we have $u(x_1,t_1)>1$ and $$ \displaystyle\max_{x\in\overline{\Omega}, T\ge t\ge0} u(x,t)=u(x_1,t_1). $$ Then $u_t(x_1,t_1)=0\ge\Delta u(x_1,t_1)$, which leads to $$ u_t-\Delta u-u(1-u)\Big|_{(x,t)=(x_1,t_1)}>0. $$ Therefore $u(x,t)\le1$. On the other hand, however, I have no idea how to prove $u(x,t)\ge0$. A similar argument for proving $u(x,t)\le1$ seems not work for the proof of $u(x,t)\ge0$. Any suggestion for ideas or proofs is welcome, thanks!

Another related question: When the homogeneous Dirichlet boundary condition $u=0$ is replaced by the homogeneous Neumann boundary condition $\frac{\partial u}{\partial \nu}=0$, the same conclusion remains true?


Consider $\displaystyle v\left(x,t\right)=u\left(x,t\right)-\frac{t}{4}$. Then you can get \begin{equation}v_t-\Delta v=u_t-\Delta u+u-u^2-\frac{1}{4}\leq0, \end{equation}which implies $v$ is subsolution of the heat equation. Hence we can apply the maximum principle to $v$ to obtain\begin{equation}\max_{\Omega}v\leq\max_{\Omega}u_0\left(x\right)\leq1.\end{equation} Thus, $u\leq1$.

On the other hand, we can set $\displaystyle w\left(x,t\right)=1-u\left(x,t\right)+\frac{t}{4}$ and $w_0\left(x\right)=1-u_0\left(x\right)$. Then we can get\begin{equation}w_t-\Delta w=w\left(w-1\right)+\frac{1}{4}\geq0,\end{equation}which means $w$ is the supersolution of the heat equation. Similarly, by using the minimum principle, we have shown\begin{equation}\min_{\Omega}w\leq\min_{\Omega}w_0\leq1.\end{equation}Thus, $u\geq0$.


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.