ABSTRACT
So far, we have only considered iterative solvers constructed on basis of a matrix splitting. In the present chapter we discuss a more sophisticated family of iterations known as Krylov subspace methods. Roughly speaking, we will use this term for any iterative solver based on approximation from a Krylov subspace. However, our main concern is the class of Krylov projection methods, which is given a precise definition in Section 3.1.
