Steady State Solution for Models with Geometric and Finite Support Activity Duration

A. Horváth

Abstract:

This paper addresses steady state solution of discrete time stochastic models in which every activity duration is given either by a geometric or a finite support distribution. Finite support distributions can be described by discrete time phase type (DPH) distributions. The behaviour of the whole stochastic model is given by a discrete time Markov chain (DTMC). The DTMC is subject to the so-called state space explosion. We present a technique for obtaining the steady state solution that alleviates this problem. The technique is based on Gaussian elimination combined with an iterative technique.

PDF

[Publications of András Horváth]