New Results on Goodness-of-Fit Tests for the Power-Law Process and Application to Software Reliability

The Power-Law process, also known as the Duane model, is a nonhomogeneous Poisson process which is widely used in reliability growth modeling. Goodness-of-fit tests for this process have been proposed by Crow, Rigdon, Klefsjö–Kumar and Park–Kim. In this paper, we propose several new tests based on chi-square, Laplace and conditional probability integral transform (CPIT) methodologies. We show that the TTT-plot test reduces to the Laplace test with Durbin's modification. We compare the power of 15 goodness-of-fit tests for the Power-Law process, and apply them to real software reliability data.