Optimal release policy of module-based software

We investigate a software reliability growth model (SRGM) based on non-homogeneous Poisson process (NHPP) for a large and complex modular software system wherein each module has a different failure rate and different failure characteristics. The main purpose of present study is to minimize the cost of software development during testing phase subject to reliability objective. Most of existing literature on the same direction deal with testing effort required during debugging phase. In this paper, a unified and integrated framework is presented for modeling of SRGM by assuming different amount of testing effort and different types of faults. The observation corresponding to isolation and removal phases in each module are considered. The model incorporates the testing effort function governed by a Weibull distribution. The optimal release policy based on cost, reliability and testing efforts is suggested. Numerical results are given to illustrate the validity of analytical results. Some special cases have also been deduced by setting appropriate parameters so as to match with the exiting results.