Optimal Release Time Determination in Intuitionistic Fuzzy Environment Involving Randomized Cost Budget for SDE-Based Software Reliability Growth Model

Abstract

To enhance the accuracy of estimating the reliability of the software, it becomes essential to count upon the proportion of faults that are predominantly dormant due to fault masking. In addition to that, the software managers and developers primarily aim at obtaining the date of product release by attaining a compromise between maximizing the reliability level and minimizing the product development cost. Keeping in mind the above requirements for reliability estimation, a stochastic differential equation-based software reliability growth model (SRGM) is developed for open-source software. Multi-release of the SRGM is formulated in detail taking into consideration the effect of bug classification. An innovative safe technique called Rx is deployed which is capable of recovering programs from both deterministic and indeterministic bugs. The optimal date of product release is obtained via a parabolic multi-objective chance-constrained nonlinear optimization problem in the parabolic intuitionistic fuzzy set environment. The optimization model framed to determine the time of product release is critically analyzed from both optimistic and pessimistic viewpoints of the developer. The results manifest the better performance of the model proposed in comparison with some preexisting SRGMs reported in the literature. Moreover, the exemplified methodology corresponding to the attainment of the optimal date for product release is well established as it is truly capable of taking into consideration, the level of tolerance of the developer pertaining to variations in both reliability and cost.