An NHPP-based SRM incorporating dependent failure mechanisms with a fault detection rate function of Pareto type Ⅱ

Software reliablility models (SRMs) based on a non-homogeneous Poisson process (NHPP) are commonly used to illustrate the process of identifying and eliminating failures throughout software testing. However, many traditional NHPP-based SRMs operate under two limiting assumptions: independence among software failures and exponentially decaying fault detection. To overcome these limitations, this paper introduces a novel NHPP-based SRM that incorporates an explicit dependent failure mechanism and a fault detection rate derived from the Pareto type II distribution. The fault detection rate is based on the hazard function of the Pareto type II distribution, leading to a hyperbolic, decreasing detection structure characterized by a long tail, which more accurately reflects the gradual decline in test effectiveness over time. Within the proposed framework, we derive analytical expressions for the mean value function and estimate model parameters using least squares. The performance of the proposed model is assessed using real software failure datasets and is compared against several existing NHPP-based SRMs.