In incremental software development approaches, the product is developed in various releases. In each release, a set of requirements is proposed for the development. Usually, due to lack of funds, lack of time and dependency between requirements, there is no possibility to develop all the required requirements. There are two conflicting objectives for choosing an optimal subset of the requirements: increasing customer satisfaction and reducing development costs. This problem is known as the next release problem (NRP) and is categorized as an NP-hard problem. Unlike the standard version of the NRP, we formulate this problem as a restricted multi-objective optimization problem. There exist metaheuristic algorithms for solving this problem performed as serials. In this paper, we introduce a parallel algorithm based on the master–slave model in order to improve the quality of the solutions. Based on the criteria of multi-objective problems, the quality of the obtained solution is compared with several metaheuristic algorithms. Two scenarios and two different datasets are used for experiments. Results indicate that the proposed method in the first scenario would highly improve the quality of solutions. Moreover, the method reduces execution time significantly through improvement in the quality of the solution in the second scenario.

中文翻译：

---

用于软件需求优化的并行多目标人工蜂群算法

在增量软件开发方法中，产品是在不同的版本中开发的。在每个版本中，都会为开发提出一组要求。通常，由于缺乏资金、缺乏时间以及需求之间的依赖性，不可能开发出所有必需的需求。选择需求的最佳子集有两个相互矛盾的目标：提高客户满意度和降低开发成本。这个问题被称为下一个版本问题 (NRP)，并被归类为 NP-hard 问题。与 NRP 的标准版本不同，我们将此问题表述为受限制的多目标优化问题。存在元启发式算法来解决作为串行执行的这个问题。在本文中，我们引入了一种基于主从模型的并行算法，以提高解决方案的质量。基于多目标问题的标准，将获得的解决方案的质量与几种元启发式算法进行比较。两个场景和两个不同的数据集用于实验。结果表明，在第一种情况下所提出的方法将大大提高解决方案的质量。此外，该方法通过提高第二场景中解决方案的质量显着减少了执行时间。结果表明，在第一种情况下所提出的方法将大大提高解决方案的质量。此外，该方法通过提高第二场景中解决方案的质量显着减少了执行时间。结果表明，在第一种情况下所提出的方法将大大提高解决方案的质量。此外，该方法通过提高第二场景中解决方案的质量显着减少了执行时间。