ABSTRACT

Finite element methods for eigenvalue problems lead to matrix eigenvalue problems. They are usually generalized eigenvalue problems, which are large and sparse. There exist many algorithms in numerical linear algebra for matrices with different properties. It is always helpful to know these properties before we choose the algebraic eigenvalue solver. Excellent books on matrix computation include, for example, [132, 221, 235]. See also the survey paper [133] and references therein. The material in this chapter is classical and can be found in [132, 79, 221].