The conjugate gradient method for queueing networks

Abstract

There are several numerical techniques to compute the steady-state probabilities for the non-Jackson queueing networks. These methods do not always converge when the coefficient matrix is a nonsingular irreducible M-matrix. In this paper we use the conjugate gradient method to compute the steady-state probabilities for the non-Jackson queueing networks. The power of the conjugate gradient is exploited by a given example of such networks. © 1995.