Software Requirements Optimization Using Multi-Objective Quantum-Inspired Hybrid Differential Evolution

Abstract

Multi-Objective Next Release Problem (MONRP) is an important software requirements optimization problem in Search-Based Software Engineering. As the customer requirements varies from time to time, often software products are required to incorporate these changes. It is a hard task to optimize the requirements from a large number of candidates, for the accomplishment of the business goals and at the same time, the satisfaction of the customers. MONRP identifies a set of requirements to be included in the next release of the product, by minimizing the cost in terms of money or resources, and maximizing the number of customers to get satisfied by including these requirements. The problem is multi-objective in nature and the objectives are conflicting objectives. The problem is NP-hard and since it cannot be solved effectively and efficiently by traditional optimization techniques especially for large problem instances, Metaheuristic Search and Optimization Techniques are required. Since MONRP has wide applicability in software companies and manufacturing companies, there is a need for efficient solution techniques especially for the large problem instances. Therefore, this paper presents a Multi-objective Quantum-inspired Hybrid Differential Evolution (MQHDE) for the solution of MONRP which combines the strengths of Quantum Computing, Differential Evolution and Genetic Algorithm. The features of MQHDE help in achieving consistency in performance in terms of convergence to Pareto-optimal front, good spread among the obtained Pareto-optimal front solutions, faster convergence and obtaining relatively more number of solutions along the Pareto-optimal front. The performance of MQHDE is tested on six benchmark problems using Spread and HyperVolume metrics. The comparison of the obtained results indicates consistent and superior performance of MQHDE over the other methods reported in the literature.