An Effective Approach to Identifying Optimal Software Reliability Allocation with Consideration of Multiple Constraints

摘要

After a software reliability goal is set for safety-critical or mission-critical software system development, it is necessary to allocate software reliability goals for individual subsystems or components so that they meet the system reliability goal. This reliability allocation also helps to allocate test resources such as cost and schedule effectively. In order to achieve these advantages, it needs to identify the optimal software reliability allocation for multiple objectives, which maximize software reliability and minimize cost and schedule. However, it is difficult to identify an optimal software reliability allocation because reliability, cost, and schedule conflict with each other. Currently, there are several software reliability allocation techniques, but schedule planning and effective resource use are unavailable because they do not support multi-objective optimization. This paper proposes an effective software reliability allocation technique that supports multi-objective optimization. The proposed approach uses a multi-objective genetic algorithm (NSGA-II) for the multi-objective optimization. This approach assists schedule planning as well as effective resources allocation. A case study is presented to show the strength of the proposed approach. Our approach helps project manager to identify well-balanced strategies with consideration of reliability, cost, and schedule in the software reliability allocation phase.