Model-reduction of nonlinear circuits using Krylov-space techniques

Abstract

A new algorithm based on Krylov subspace methods is proposed for model-reduction of large nonlinear circuits. Reduction is obtained by projecting the original system described by nonlinear differential equations into a subspace of a lower dimension. The reduced model can be simulated using conventional numerical integration techniques. Significant reduction in computational expense is achieved as the size of the reduced equations is much less than that of the original system.