The invariance principle for nonlinear Fokker--Planck equations

TL;DR

This paper investigates the long-term behavior of solutions to nonlinear Fokker-Planck equations using the La Salle invariance principle, establishing the existence of equilibrium states and the properties of their omega-limit sets in an infinite-dimensional setting.

Contribution

It applies the La Salle invariance principle to nonlinear semigroups in Banach spaces to analyze the omega-limit sets and equilibrium states of nonlinear Fokker-Planck equations, extending previous results.

Findings

Existence of an equilibrium state for the nonlinear Fokker-Planck equation.

The omega-limit set is non-empty, compact, and invariant under the solution operator.

In the nondegenerate case, solutions converge to a unique stationary state.

Abstract

One studies here, via the La Salle invariance principle for nonlinear semigroups in Banach spaces, the properties of the $\omega$-limit set $\omega(u_0)$ corresponding to the orbit $\gamma(u_0)=\{u(t,u_0);\ t\ge0\}$, where $u=u(t,u_0)$ is the solution to the nonlinear Fokker-Planck equation $$\begin{array}{l} u_t-\Delta\beta(u)+{\rm div}(Db(u)u)=0\ \mbox{ in }(0,\infty)\times\mathbb{R}^d,\\ u(0,x)=u_0(x),\ \ x\in\mathbb{R}^d,\quad u_0\in L^1(\mathbb{R}^d),\ d\ge3.\end{array}$$ Here, $\beta\in C^1(\mathbb{R})$ and $\beta'(r)>0$, $\forall r\ne0$. Moreover, $\beta$ is a sublinear function, possibly degenerate in the origin, $b\in C^1(\mathbb{R})$, $b$ bounded, $b\ge b_0\in(0,\infty),$ $D$ is bounded such that $D=-\nabla\Phi$, where $\Phi\in C(\mathbb{R}^d)$ is such that $\Phi\ge1,$ $\Phi(x)\to\infty$ as $|x|\to\infty$ and satisfies a condition of the form $\Delta\Phi-\alpha|\nabla\Phi|^2\le0$, a.e. on $\mathbb{R}^d$. The main conclusion is that the equation has an equilibrium state and the set $\omega(u_0)$ is a non-empty, compact subset of $L^1(\mathbb{R}^d)$ while, for each $t\ge0$, the operator $u_0\to u(t,u_0)$ is an isometry on $\omega(u_0)$. In the nondegenerate case $0<\gamma_0\le\beta'\le\gamma_1$ studied in [Barbu, R\"ockner: arXiv:1808.10706], it follows that $\lim\limits_{t\to\infty}S(t)u_0=u_\infty$ in $L^1(\mathbb{R}^d)$, where $u_\infty$ is the unique bounded stationary solution to the equation.