Reduced-order modeling techniques based on Krylov subspaces and their use in circuit simulation (Q4952744)

The aim of this paper is to survey the main ideas involved in reduced-order modelling techniques based on Krylov subspaces, and to describe the use of these methods within the field of large-scale time-invariant linear dynamical systems that arise in the simulation of electronic circuits. The contents of the ten sections is as follows: Section 2 formulates the reduced-order modelling problem for linear dynamical systems. Section 3 describes how linear dynamical systems arise in the process of circuit simulation. Section 4 defines the notions of Krylov and block Krylov subspaces for single and, respectively, multiple starting vectors. The connection of Krylov subspaces to moment matching for SISO (single-input-single-output) linear dynamic systems is discussed. Section 5 exposes the classical Lanczos process and its extensions to multiple starting vectors. Section 6 discusses the use of the Lanczos methods for reduced-order modelling. Section 7 introduces the Arnoldi process, while Section 8 surveys reduced-order modelling techniques that are based on the Arnoldi process. Section 9 examines the use of reduced-order modelling for circuit-noise computations, and shows how noise-type transfer functions can be rewritten such that the reduced-order modelling techniques are applied to linear dynamical systems. Concluding remarks are enclosed into the final Section 10. For the entire collection see [Zbl 0934.00020].