Non-homogeneous Markov Process Modeling for Software Reliability Assessment

摘要

In this paper, we focus on non-homogeneous Markov processes, which are generalizations of the well-known non-homogeneous Poisson processes, and compare two software reliability models (SRMs) classified into a generalized binomial process (GBP) and a generalized Polya process (GPP). GBP and GPP can be characterized respectively as a Markov death process and a Markov birth process, with state and time dependent transition rates. Through numerical examples with the fault count data observed in actual software development projects, we compare the two SRMs in terms of the goodness-of-fit performances.