Non‐homogeneous Poisson Process Models for Software Reliability
Steven E. Rigdon · 2018
The non-homogeneous Poisson process (NHPP) has long been used as a model for the reliability of repairable hardware systems. This chapter covers the NHPP and some of its properties. It also covers the piecewise exponential (PEXP) models. The chapter describes generalizations, or hybrids, of these models. There are two basic paradigms for collecting data from software testing. One involves the recording of the exact time of each failure, or equivalently, the times between failures. The other is to record, for several time intervals, the time tested along with the number of failures in that interval. The chapter discusses statistical inference for both approaches. It also discusses two examples: one involves the recording of the times between failure, and the other involves counting how many failures fell in each interval. The chapter describes some other models that have been applied to software reliability.