Optimal Testing Resource Allocation for Modular Software Based on a Software Reliability Growth Model: A Dynamic Programming Approach

In this paper, two problems of optimal resource allocation to modules during testing phase are studied: (1) maximization of the number of faults removed when the amount of testing-effort is fixed, and (2) maximization of the number of faults removed satisfying a certain percentage of initial faults to be removed with a fixed amount of testing-effort. These optimization problems are formulated as nonlinear programming problems (NLPP), which are modeled by a software reliability growth model based on a non-homogeneous Poisson process which incorporated the exponentiated Weibull testing-effort functions. A solution procedure is then developed using dynamic programming technique to solve the NLPPs. Finally, numerical examples illustrate the solution procedure and the results are compared with that of Kapur et al.[1].