Computational methods for nonlinear elliptic eigenvalue problems (Q758156)

Eigenvalue problems of the form \(-\Delta u=\lambda F_ u(x,u)\) in D, \(u=0\) on \(\partial D\), are considered where D is a bounded domain of class \(C^ 1\) in \(R^ N\), \(N=1,2,3,\) and F is a \(C^ 1\)-function of \(x\in D\) and u. A transforming function g is used to transform the problem into a sequence of quadratic variational problems. Three examples of such transforming functions are discussed, the first one working only for F being convex in u. Tests with \(F(x,u)=-\cos (u)\) in the other cases demonstrate the use of the method in the nonconvex case. The main theoretical result - a kind of global convergence theorem - states that the iterates \(u_ n\) generated by the method approach the set \(\Omega_ R\) of generalized eigenfunctions with fixed norm R in the Sobolev space \(H^ 1_ 0(D)\), a subsequence of \(\{u_ n\}\) converges in \(H^ 1_ 0\) to a generalized eigenfunction \(u^*\).