A Mixed Poisson Process and Empirical Bayes Estimation Based Software Reliability Growth Model and Simulation

摘要

A new Software Reliability Model based on a Mixed Poisson process where the failure rate follows an Inverse Gaussian distribution is proposed. By using the Empirical Bayes estimate of the failure rate, our estimate depends just on the number of failures and the total past time, it is not necessary to know the exact instants of time were the failures have occurred. We simulate the failure detection process with exponential waiting times randomly generated where the failure rate is constantly updated with the Empirical Bayes estimate. Results of simulations are compared with two real projects.