Modeling a System of Softwares Under Imperfect Debugging

This paper considers a series system of p softwares where the failure of a software follows a modified Jelinski-Moranda model. In this model, the debugging scheme of a software is imperfect in the sense that the number of bugs introduced to the software that caused system failure is a random variable. Moreover, tasks are assigned to the system with the task completion times assumed to be random. The main interest is to estimate the parameters that describe the failure process of each software as well as the system reliability. Finally, the finite-sample properties of these estimators are investigated via a simulation