Compactness results for Sign-Changing Solutions of critical nonlinear elliptic equations of low energy

Hussein Cheikh Ali, Bruno Premoselli · 2024

Read Compactness results for Sign-Changing Solutions of critical nonlinear elliptic equations of low energy by Hussein Cheikh Ali, Bruno Premoselli (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-00817.txt.

Introduction

Hussein Cheikh Ali and Bruno Premoselli wrote a 2024 math paper on solutions of a hard equation from the study of curved surfaces and fields. The equation has a sharp growth term that often makes solutions blow up at points. The authors study solutions that change sign, at the lowest such energy level.

Their main result says the set of those low-energy solutions stays tightly packed in spaces of three to five dimensions. In higher dimensions the same holds when a key term never hits zero. The proof gives a fresh global picture of how blow-up can occur, right up to the edge of the region. The results extend older work on this classic problem.

Catalog record

CREATORHussein Cheikh Ali, Bruno Premoselli
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE118 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-00817

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Compactness results for Sign-Changing Solutions of critical nonlinear elliptic equations of low energy Thanks: The second author was supported by the Fonds Thélam, an ARC Avancé 2020 grant and an EoS FNRS grant Hussein Cheikh Ali, Bruno Premoselli Abstract . Let Ω \Omega be a bounded, smooth connected open domain in ℝ n \mathbb{R}^{n} with n ≥ 3 n\geq 3 . We investigate in this paper compactness properties for the set of sign-changing solutions v ∈ H 0 1 ​ ( Ω ) v\in H^{1}_{0}(\Omega) of (*) { − Δ ​ v + h ​ v = | v | 2 ∗ − 2 ​ v in ​ Ω , v = 0 on ​ ∂ Ω \left\{\begin{aligned} -\Delta v+hv&=\left|v\right|^{2^{*}-2}v&\hbox{ in }\Omega,\\ v&=0&\hbox{ on }\partial\Omega\end{aligned}\right. where h ∈ C 1 ​ ( Ω ¯ ) h\in C^{1}(\overline{\Omega}) and 2 ∗ := 2 ​ n / ( n − 2 ) 2^{*}:=2n/(n-2) . Our main result establishes that the set of sign-changing solutions of ( * ) at the lowest sign-changing energy level is unconditionally compact in C 2 ​ ( Ω ¯ ) C^{2}(\overline{\Omega}) when 3 ≤ n ≤ 5 3\leq n\leq 5 , and is compact in C 2 ​ ( Ω ¯ ) C^{2}(\overline{\Omega}) when n ≥ 7 n\geq 7 provided h h never vanishes in Ω ¯ \overline{\Omega} . In dimensions n ≥ 7 n\geq 7 our results apply when h > 0 h>0 in Ω ¯ \overline{\Omega} and thus complement the compactness result of [ 16 ] . Our proof is based on a new, global pointwise description of blowing-up sequences of solutions of ( * ) that holds up to the boundary. We also prove more general compactness results under perturbations of h h . 1. Introduction 1.1. Statement of the results Let \Omega\subset\mathbb{R}^{n} be a smooth bounded connected open set in \mathbb{R}^{n} , n\geq 3 , h\in C^{1}(\overline{\Omega}) and 2^{*}:=2n/(n-2) . In this paper we investigate solutions v\in H^{1}_{0}(\Omega) of (1.1) \left\{\begin{aligned} -\Delta v+hv&=\left|v\right|^{2^{*}-2}v&\hbox{ in }\Omega,\\ v&=0&\hbox{ on }\partial\Omega.\end{aligned}\right. Here and in the sequel, we let \mathbb{\|}\cdot\mathbb{\|}_{p} be the usual norm of L^{p}(\Omega) for 1\leq p\leq\infty , and H_{0}^{1}(\Omega) be the completion of C_{c}^{\infty}(\Omega) with respect to the norm \mathbb{\|}v\mathbb{\|}_{H_{0}^{1}}^{2}:=\int_{\Omega}|\nabla v|^{2}\,dx. For simplicity we will assume throughout this paper that -\Delta+h is coercive, that is, that there exists C>0 such that \int_{\Omega}\left(|\nabla v|^{2}+hv^{2}\right)\,dx\geq C\int_{\Omega}|\nabla v|^{2}\,dx\hbox{ for all }v\in H_{0}^{1}(\Omega). Under this assumption, the existence of positive solutions of ( 1.1 ) is very well-understood. We let (1.2) I_{h}(\Omega):=\inf_{v\in H_{0}^{1}(\Omega)\backslash\{0\}}\frac{\int_{\Omega}\left(|\nabla v|^{2}+hv^{2}\right)\,dx}{\left(\int_{\Omega}|v|^{2^{*}}\,dx\right)^{\frac{2}{2^{*}}}}. Brézis-Nirenberg [ 8 ] proved that when n\geq 4 positive ground states attaining ( 1.2 ) exist if and only h<0 somewhere in \Omega . When n=3 , Druet [ 17 ] proved that positive ground states attaining ( 1.2 ) exist if only if m_{h}>0 somewhere in \Omega , where m_{h} is the so-called mass-function of the operator -\Delta+h . This function is defined as follows: let G_{h} be the Green’s function for -\Delta+h with Dirichlet boundary conditions in \Omega . Then, when n=3 , we have G_{h}(x,y)=\frac{1}{4\pi|x-y|}+g_{h}(x,y)\hbox{ for all }y\in\Omega\backslash\{x\} for some g_{h}\in C^{0,1}(\overline{\Omega}\times\overline{\Omega}) , and we define m_{h}(x)=g_{h}(x,x) . Under these assumptions, [ 8 ] and [ 17 ] also prove that we have I_{h}(\Omega)<K_{n}^{-2} , where (1.3) K_{n}^{-2}:=\inf_{v\in C_{c}^{\infty}(\mathbb{R}^{n})\backslash\{0\}}\frac{\int_{\mathbb{R}^{n}}|\nabla v|^{2}\,dx}{\left(\int_{\mathbb{R}^{n}}|v|^{2^{*}}\,dx\right)^{\frac{2}{2^{*}}}} is the optimal constant in Sobolev’s inequality in \mathbb{R}^{n} . An explicit expression of K_{n} can be found in [ 1 , 53 ] . It is simple to see that if v\in H^{1}_{0}(\Omega) attains I_{h}(\Omega) then (1.4) \int_{\Omega}|v|^{2^{*}}\,dx=I_{h}(\Omega)^{\frac{n}{2}}<K_{n}^{-n}. The existence of sign-changing solutions for problem ( 1.1 ) has also attracted a lot of attention. Existence results for a general function h\in C^{1}(\overline{\Omega}) are in [ 3 ] . When h\equiv-\lambda , for \lambda\in(0,\lambda_{1}) , equation ( 1.1 ) is the so-called Brézis-Nirenberg problem: (1.5) \left\{\begin{aligned} -\Delta v-\lambda v&=\left|v\right|^{2^{*}-2}v&\hbox{ in }\Omega\\ v&=0&\hbox{ on }\partial\Omega,\end{aligned}\right. for which existence results have been obtained in [ 11 , 9 , 24 , 50 , 16 , 14 , 49 ] . The existence of a sign-changing solution of least-energy (among all sign-changing solutions) for ( 1.5 ) when \lambda\in(0,\lambda_{1}) – the range in which -\Delta-\lambda is coercive – was proven in [ 10 ] when n\geq 6 (see also [ 13 ] for a new proof) while it was proven in [ 48 , 54 ] when n=4,5 . The existence of least-energy sign-changing solutions for ( 1.5 ) is not yet known when n=3 . In this paper we focus on compactness properties for solutions of ( 1.1 ). We let (h_{\alpha})_{\alpha\in\mathbb{N}} be a sequence of C^{1} functions that converge to h in C^{1}(\overline{\Omega}) and we let (v_{\alpha})_{\alpha\in\mathbb{N}} be a sequence of solutions in H^{1}_{0}(\Omega) of (1.6) \left\{\begin{aligned} -\Delta v_{\alpha}+h_{\alpha}v_{\alpha}&=\left|v_{\alpha}\right|^{2^{*}-2}v_{\alpha}&\hbox{ in }\Omega,\\ v_{\alpha}&=0&\hbox{ on }\partial\Omega\end{aligned}\right. satisfying \limsup_{\alpha\to+\infty}\|v_{\alpha}\|_{H^{1}_{0}}<+\infty . We will say that (v_{\alpha})_{\alpha} is sign-changing if, for any \alpha , (v_{\alpha})_{+}=\max(v_{\alpha},0) and (v_{\alpha})_{-}=-\min(v_{\alpha},0) are both nonzero. We investigate under which assumptions on h the sequence (v_{\alpha})_{\alpha\in\mathbb{N}} converges in a strong topology. Our main result answers this question when (v_{\alpha})_{\alpha\in\mathbb{N}} has minimal energy: Theorem 1.1 . Let \Omega be a smooth bounded connected domain of \mathbb{R}^{n} , n\geq 3 , and (h_{\alpha})_{\alpha\in\mathbb{N}} be a sequence that converges in C^{1}(\overline{\Omega}) towards h . Assume that -\Delta+h is coercive and that I_{h}(\Omega)<K_{n}^{-2} . Let (v_{\alpha})_{\alpha\in\mathbb{N}}\in H_{0}^{1}(\Omega) be a sequence of solutions of ( 1.6 ) such that (1.7) \limsup_{\alpha\to+\infty}\int_{\Omega}|v_{\alpha}|^{2^{*}}\,dx\leq K_{n}^{-n}+I_{h}(\Omega)^{\frac{n}{2}} and assume that one of the following assumptions is satisfied: • either n\in\{3,4,5\} and, for all \alpha\geq 0 , v_{\alpha} is sign-changing, or • n\geq 7 and h\neq 0 at every point in \overline{\Omega} . Then, up to a subsequence, (v_{\alpha})_{\alpha\in\mathbb{N}} strongly converge in C^{2}(\overline{\Omega}) to a non-zero solution of ( 1.1 ). Recall that I_{h}(\Omega) is defined in ( 1.2 ). In the particular case where h_{\alpha}\equiv h , Theorem 1.1 implies the following compactness result for solutions of ( 1.1 ): Corollary 1.1 . Let \Omega be a smooth bounded connected domain of \mathbb{R}^{n} , n\geq 3 , and let h\in C^{1}(\overline{\Omega}) be such that -\Delta+h is coercive and I_{h}(\Omega)<K_{n}^{-2} . • Assume that n\in\{3,4,5\} . There exists \varepsilon=\varepsilon(n,\Omega)>0 such that the set of sign-changing solutions v of ( 1.1 ) satisfying \int_{\Omega}|v|^{2^{*}}\,dx\leq K_{n}^{-n}+I_{h}(\Omega)^{\frac{n}{2}}+\varepsilon is precompact in the C^{2}(\overline{\Omega}) -topology. • Assume that n\geq 7 and h\neq 0 in \overline{\Omega} . There exists \varepsilon=\varepsilon(n,h,\Omega)>0 such that the set of solutions v of ( 1.1 ) satisfying \int_{\Omega}|v|^{2^{*}}\,dx\leq K_{n}^{-n}+I_{h}(\Omega)^{\frac{n}{2}}+\varepsilon is precompact in the C^{2}(\overline{\Omega}) -topology. The energy bound ( 1.7 ) is very natural when investigating sign-changing solutions of ( 1.1 ). Solutions of ( 1.6 ) satisfying ( 1.7 ) exist: the least-energy sign-changing solutions of ( 1.5 ) constructed in [ 10 , 54 ] , for instance, satisfy \int_{\Omega}|v|^{2^{*}}\,dx<K_{n}^{-n}+I_{-\lambda}(\Omega)^{\frac{n}{2}} . A simple application of Struwe’s [ 51 ] celebrated compactness result (see also [ 10 , Lemma 3.1] ) shows that if a sequence (v_{\alpha})_{\alpha\in\mathbb{N}} of solutions of ( 1.6 ) changes sign and satisfies \lim_{\alpha\to+\infty}\|v_{\alpha}\|_{\infty}=+\infty (we will say in this case that (v_{\alpha})_{\alpha\in\mathbb{N}} blows-up ), then \int_{\Omega}|v_{\alpha}|^{2^{*}}\,dx\geq K_{n}^{-n}+I_{h}(\Omega)^{\frac{n}{2}}+o(1) as \alpha\to+\infty . The threshold K_{n}^{-n}+I_{h}(\Omega)^{\frac{n}{2}} is therefore the direct counterpart, for sign-changing solutions, of the minimal energy threshold K_{n}^{-n} that ensures the existence of positive ground state solutions in ( 1.4 ). In this respect, Theorem 1.1 and Corollary 1.1

Read the complete text