Maximum principle for nonlinear parabolic equations (Q1991897): Difference between revisions

In this paper, the author gives some maximum principles for solutions of some nonlinear parabolic equations. The author considers first, the parabolic equation \[ \mathcal{L}u-u_{t}=f(x,t,u,Du),\text{ in }\Omega\cup\gamma\Omega\tag{1} \] where {\parindent=6mm \begin{itemize}\item[{\(\bullet\)}] \(\Omega\subset\mathbb{R}^n\times(0,\infty)\) is an unbounded open set, \(\partial\Omega=\Gamma(\Omega)\cup\gamma(\Omega)\), \(\Gamma(\Omega)\) is the proper boundary of \(\Omega\) and \(\gamma(\Omega)\) is the top side of the boundary; \item[{\(\bullet\)}] \(\mathcal{L}u=\sum_{i,j=1}^na_{ij}(x,t,u)\partial_i\partial_j u+\sum_{i=1}^nb_{i}(x,t,u)\partial_i u\) is a differential operator with symmetric coefficients and \(\sum_{i,j=1}^na_{ij}(x,t,s)\xi_{i}\xi_{j}>0\), for all \((x,t)\in\Omega\cup\gamma(\Omega)\), \(s\in\mathbb{R}\setminus\{0\}\) and \(\xi=(\xi_1,\dots,\xi_{n})\in\mathbb{R}^n\setminus\{0\}\). \end{itemize}} A function \(u\) is called a solution of the equation (1), if it is continuous in \(\Omega\cup\gamma(\Omega)\), possesses two continuous derivatives on this set with respect to \(x\) and one continuous derivative in \(t\), and satisfies equation (1). It is well known that if the domain \(\Omega\) is bounded, then the maximum principle \[ \sup_{\Omega}|u|\leq\lim_{(x,t)\to\Gamma(\Omega)}\sup|u(x,t)|\tag{2} \] holds for the solution of equation (1). The author gives conditions in which (2) holds for an unbounded domain. Theorem 2.1. provides conditions in which (2) holds for any solution of (1). After that, the author considers the equation \[ \mathcal{L}u-\partial_{t}u=q(x)g(|u|)\operatorname{sign} u,\text{ in }\Omega\cup\gamma\Omega,\tag{3} \] where \(q:\Omega\cup\gamma\Omega\to[0,\infty)\), and \(g:[0,\infty)\to[0,\infty)\) is a locally bounded measurable functions such that \[ \inf_{K}g>0,\tag{4} \] for any compact \(K\subset(0,\infty)\). Theorem 2.2. provides conditions in which (2) holds for any solution of (3). Next, the author considers the nonlinear Fokker-Planck equation \[ \operatorname{div}(u^{\alpha}Du)+\beta(x)Du-u_{t}=q(x)g(u),\text{ in }\Omega\cup\gamma\Omega,\tag{5} \] where \(\alpha>0\) is a real constant, \(\beta:\Omega\cup\gamma\Omega\to\mathbb{R}^n\), \(q:\Omega\cup\gamma\Omega\to[0,\infty)\), \(g:(0,\infty)\to[0,\infty)\), are locally bounded measurable functions such that (4) holds. Because, in general, \(u\) from (5) is interpreted as thermodynamic temperature, it is positive. Therefore (2) becomes \[ \sup_{\Omega}u\leq\lim_{(x,t)\to\Gamma(\Omega)}\sup u(x,t).\tag{6} \] Theorem 2.3 provides conditions in which (6) holds for any solution of (5). Theorem 2.4--2.6 can be regarded as the analogues of Theorems 2.1--2.3 for differential inequalities.