Asymptotic behavior for the Brezis-Nirenberg problem. The subcritical perturbation case

TL;DR

This paper investigates the asymptotic behavior, profile, and blow-up rate of positive least energy solutions to a perturbed Brezis-Nirenberg problem as the perturbation parameter approaches zero, revealing dependence on domain geometry and the exponent q.

Contribution

It provides a sharp asymptotic characterization and establishes uniqueness and nondegeneracy of solutions for the subcritical perturbation case, extending previous results for q=2.

Findings

Asymptotic profile and blow-up rate of solutions characterized.

Proved uniqueness and nondegeneracy under mild domain assumptions.

Dependence of blow-up rate on dimension, domain geometry, and exponent q.

Abstract

In this paper, we are concerned with the well-known Brezis-Nirenberg problem \begin{equation*} \begin{cases} -\Delta u= u^{2^*-1}+\varepsilon u^{q-1},\quad u>0, &{\text{in}~\Omega},\\ \quad \ \ u=0, &{\text{on}~\partial \Omega}, \end{cases} \end{equation*} where $\Omega\subset \mathbb R^N$ with $N\ge 3$ is a bounded domain, $q\in(2,2^*)$ and $2^*=\frac{2N}{N-2}$ denotes the critical Sobolev exponent. It is well-known (H. Br\'{e}zis and L. Nirenberg, \newblock {\em Comm. Pure Appl. Math.}, 36(4):437--477, 1983) that the above problem admits a positive least energy solution for all $\varepsilon >0$ and $q>\max\{2,\frac{4}{N-2}\}$. In the present paper, we first analyze the asymptotic behavior of the positive least energy solution as $\varepsilon\to 0$ and establish a sharp asymptotic characterisation of the profile and blow-up rate of the least energy solution. Then, we prove the uniqueness and nondegeneracy of the least energy solution under some mild assumptions on domain $\Omega$. The main results in this paper can be viewed as a generalization of the results for $q=2$ previously established in the literature. But the situation is quite different from the case $q=2$, and the blow-up rate not only heavily depends on the space dimension $N$ and the geometry of the domain $\Omega$, but also depends on the exponent $q\in(\max\{2,\frac{4}{N-2}\}, 2^*)$ in a non-trivial way.