ABSTRACT
The Lindley distribution and its modifications have garnered much attention in recent years. In lieu of the usually utilized exponential distribution, they are employed in the life data analysis of various items. Several novel software reliability models (SRMs) based on the non-homogeneous Poisson process (NHPP) are proposed in this chapter. We suppose that the epochs of software failures correspond to the record values of i.i.d. random variables distributed according to seven Lindley-type distribution functions and refer to these SRMs as infinite Lindley-type NHPP-based SRMs. In the numerical example, we evaluate the goodness of fit and predictive performances of infinite Lindley-type NHPP-based SRMs and apply them to predict quantitative software reliability by two types of failure count data observed in several actual software development projects.
