Multiplicity of solutions to the quasilinear Neumann problem in the 3-dimensional case (Q2515212)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Multiplicity of solutions to the quasilinear Neumann problem in the 3-dimensional case |
scientific article |
Statements
Multiplicity of solutions to the quasilinear Neumann problem in the 3-dimensional case (English)
0 references
31 July 2015
0 references
Let \(B_R\) be a ball in \(\mathbb{R}^3\), with radius \(R>0\), and let \(S_R\) be the boundary of \(B_R\). This paper deals with the following problem: \[ \begin{cases} -\Delta_pu+|u|^{p-2}u=0 \;\text{ in }B_R,\\ |\nabla u|^{p-2}\langle \nabla u,{\mathbf n}\rangle= |u|^{q-2}u \;\text{ on }S_R,\\ u>0 \;\text{ in }B_R,\end{cases} \] where \({\mathbf n}\) is the unit outward normal to \(S_R\), \(p\in (1,+\infty)\), \(q\in \biggl(p,\;\frac{2}{3}\biggl(\frac{3p}{3-p}\biggr)_+\biggr)\), and \(\Delta_p(u):=\)div\((|\nabla u|^{p-2}\nabla u)\) is the \(p\)-Laplacian operator. The authors study the multiplicity of nonequivalent solutions to this problem. Solutions are said nonequivalent if none of them can be obtained from each other by a rotation. The authors prove that, for each fixed \(N\in \mathbb{N}\), the problem admits at least \(N\) nonequivalent solutions provided that \(R\) is large enough. To prove this result, the authors consider the group \(\mathcal{G}_k\), \(k\in \mathbb{N}\), generated by rotation about the \(x_3\)-axis by \(\frac{2\pi}{k}\) and symmetry relative to the plane \(Ox_1x_2\), and the functional \[ J[u]=\frac{\int_{B_R}(|\nabla u|^p+|u|^p)dx }{(\int_{S_R}|u|^qdS)^{\frac{p}{q}}} \] restricted to the the set \(A\) of all functions \(u\in W_p^1(B_R)\) which are \(\mathcal{G}_k\)-invariant and such that \(\int_{S_R}|u|^qdS=1\). The authors prove that \(J_{\mid A}\) admits a local minimum \(u_R\), for \(R\) large. From the principle of symmetric criticality and the Harnack inequality, the authors infer that \(u_R\) is a solution of the problem. Then, the authors prove that, for each fixed \(N\in \mathbb{N}\), and each \(k\in \{1,\dots,N\}\), there exist \(R_k>0\) such that, for \(R>R_k\), \(u_R\) has exactly \(k\) ``peaks''. From this, the authors infer that the problem has, for any \(R>\max_{1\leq k\leq N}R_k\), at least \(N\) nonequivalent solutions. A similar result was obtained for \(n\)-dimensional balls, with \(n\neq 3\), in [\textit{A. P. Shcheglova}, Probl. Mat. Anal. 30, 121--144 (2005; Zbl 1099.35039); translation in J. Math. Sci., New York 128, No. 5, 3306--3333 (2005)].
0 references
Neumann problem
0 references
\(p\)-Laplacian
0 references
nonequivalent solutions
0 references
variational methods
0 references
concentration
0 references
0 references
0 references
0 references