Finite modeling of parabolic equations using Galerkin methods and inverse matrix approximations (Q2563555)

This paper presents a method to obtain a reduced-order model of a parabolic system using the Galerkin method and inverse matrix approximations. Bounds are given for the approximation error and reduction error. The approximation error is tied to the diagonal dominance of the matrix whose inverse is approximated. The reduction error is linked to a modal truncation type error bound. The proposed algorithm offers great advantage by the low computational complexity.