Software Requirement Prioritization using Artificial Bee Colony Algorithm

Abstract

Requirement Prioritization plays an important role in the software development. To achieve effective outcome, a variety of techniques have been employed to prioritize software requirements. There are two disagreeing aspects; maximizing the satisfaction of the customers and minimizing the costs of development, that we face while selecting an optimal set of requirements and it is categorized as NP-hard problem. If we focus on the acquiring the approval from the customer, the price of the development increases whereas when we focus on cost, there are high chances of reduction in customer satisfaction. To solve this problem of prioritization, many meta-heuristic algorithms have been used and researches on including many genetic and evolutionary algorithms. Some focus on single-objective where others develop solutions for multiple objectives. Some of the simple methods such as hierarchy methods, ranking and numerical assignments have been discussed as well. In our research, we focus on the artificial bee colony algorithm to provide effective results for requirement optimization. We have proposed a method to prioritize requirements using the clustering technique with the help of artificial bee colony algorithm. The requirements with similar importance are assembled together in same groups. Our goal is to improve the prioritization results and reduce the errors obtained in the results as much as possible while maintaining the best outcome. The algorithm is executed on selected datasets of RALIC and the results are analyzed. In this research, a benchmark study is used to compare the results of the prioritization. The experimental results show that using clustering with ABC algorithm, the prioritization of requirements has been effectively improved.