U-plot method for testing the goodness-of-fit of the Power-Law Process

This paper shows that the u-plot method can be used as a statistical procedure for testing the fit of the Power-Law-Process (PLP), also called the Duane model. The u-plot was initially presented as a graphical tool for validating software reliability models. Mathematical derivations are given here to justify the use of u-plot as a statistical goodness-of-fit test for the PLP. These results give theoretical explanations to Downs and Scott simulation results. Methods and tools are suggested to further investigations which should lead to proving that the u-plot method can be considered as a common statistical goodness-of-fit test for several reliability growth models. The u-plot test is compared to classical PLP goodness-of-fit tests for different alternative models, and appears to perform very well.