Solving Large, Sparse Eigenvalue Problems with Hamiltonian Structure
David S. Watkins
Department of Mathematics, Washington State University
Pullman, WA, USA
Abstract: A class of quadratic eigenvalue problems arising from three-dimensional elasticity will be discussed. Since the matrices are obtained from finite element discretizations, they are large and sparse.
Even though the underlying problem is static, not dynamic, the resulting eigenvalue problem has Hamiltonian structure, which should be exploited if the problem is to be solved efficiently.
Three structure-preserving approaches will be presented, along with some encouraging preliminary numerical results.