Maximum Principle and principal eigenvalue in unbounded domains under general boundary conditions

TL;DR

This paper explores the relationship between the Maximum Principle and the principal eigenvalue for elliptic operators in unbounded domains, covering various boundary conditions and establishing new theoretical results.

Contribution

It provides a unified framework linking the Maximum Principle to the principal eigenvalue under general boundary conditions, including new results for classical cases.

Findings

Positivity of the principal eigenvalue is necessary and sufficient for the Maximum Principle.

Zero principal eigenvalue implies the Critical Maximum Principle under certain conditions.

Counterexamples show limitations and clarify misconceptions about eigenvalue simplicity.

Abstract

This paper investigates the link between the Maximum Principle and the sign of the (generalized) principal eigenvalue for elliptic operators in unbounded domains. Our approach covers the cases of Dirichlet, Neumann, and (indefinite) Robin boundary conditions and treat them in a unified way. For a certain class of elliptic operators (including the class of selfadjoint operators), we establish that the positivity of the principal eigenvalue is a necessary and sufficient condition for the validity of the Maximum Principle. If the principal eigenvalue is zero, no general answer holds; instead, under a natural condition on the domain's size at infinity, we show that the operator satisfies what we call the Critical Maximum Principle. We also address the question of the simplicity of the principal eigenvalue, and a series of counterexamples is proposed to disprove some possible misconceptions. Our main results are new even for the more classical cases of Dirichlet boundary conditions and selfadjoint operators.