Periodic solutions of superlinear beam and membrane equations with perturbations from symmetry (Q5935346): Difference between revisions

It is well known in variational methods that when the functional under study is equivariant, we get multiplicity results with infinitely many solutions. A natural question that stems then is wether it is really the symmetry that is responsible for these spectacular multiplicity results. Would some results persist if some perturbation is injected? This stability problem is quite old, and mathematicians were interested in the study of the effects of breaking the symmetry by introducing ``small perturbations'' since the apparition of the Ljusternik-Schnirelman theory. Among the recent prospections in this direction, we can cite Bahri and Berestycki [Trans. Am. Math. Soc. 267, 1--32 (1981; Zbl 0476.35030)], Struwe [Manuscr. Math. 32, 335--364 (1980; Zbl 0456.35031)], Rabinowitz [Trans. Am. Math. Soc. 272, 753--769 (1982; Zbl 0589.35004), Reg. Conf. Ser. Math. 65 (1986; Zbl 0609.58002)], Bahri and Lions [Commun. Pure Appl. Math. 41, 1027--1038 (1988; Zbl 0645.58013)], Clapp [Comment. Math. Helv. 71, No. 4, 570--593 (1996; Zbl 0873.58018)] for elliptic boundary value problems, Bahri and Berestycki [Acta Math. 152, 143-197 (1984; Zbl 0592.70027)], Mawhin and Willem [Applied Mathematical Sciences, 74. New York etc.: Springer-Verlag. (1989; Zbl 0676.58017)], Long [Math. Z. 203, No. 3, 453--467 (1990; Zbl 0687.34034], Bolle [J. Differ. Equations 152, No. 2, 274--288 (1999; Zbl 0923.34025)], Bolle, Ghoussoub, and Tehrani, [Manuscr. Math. 101, No. 3, 325-350 (2000; Zbl 0963.35001)] for Hamiltonian systems and Tanaka [Ann. Mat. Pura Appl., IV. Ser. 162, 43--76 (1992; Zbl 0801.35087), Trans. Am. Math. Soc. 307, 615--645, (1988; Zbl 0696.35113)] for wave equations. This paper is another contribution in this effort. The authors consider the problem of perturbation from symmetry for the so-called ``generalized membrane equation'' \((P_{n})\) \(u_{tt}+\Delta_{x}^{2} u = f(x,t,u)\) for \((x,t)\in \Omega_{n}\times\mathbb R=(0,\pi)^{n}\times\mathbb R\), \(u(x,t)=D_{ii}u(x,t)=0\) for \((x,t)\in \partial \Omega_{n}\times \mathbb R\), \(i=1,\ldots,n\), \(u(x,t)=u(x,t+2\pi)\) for \((x,t)\in \Omega_{n}\times\mathbb R\), where \(\Delta_{x}=\sum_{i=1}^{n}D_{ii}\), \(D_{ii}=\partial^{2}/\partial x_{i}^{2}\), \(f(x,t,\xi)=g(x,t,\xi) + p(x,t,\xi)\) is \(2\pi\)-periodic in \(t\), \(g\) is odd in \(x\) but not \(p\). The authors point out that, although \((P_{n})\) looks similar to the wave problem studied by Tanaka, the energy associated to \((P_{n})\) does not satisfy the Palais-Smale condition and the kernel of the quadratic part of the energy is infinite dimensional and more complicated than the kernel of the wave operator. They prove existence and multiplicity of solutions for \((P_{n})\) using a specific abstract critical point theorem developed for this aim.