Selecting a set of requirements to implement in the next software release is an NP-Hard problem known as NRP. We propose multi-objective versions of grey wolf optimizer and whale optimization algorithm for solving bi-objective NRP. We used these two algorithms and three other evolutionary algorithms to solve NRP problem instances from four datasets. The cost-to-score ratio and the roulette wheel are used to satisfy constraints of the NRP problem. We compare obtained Pareto fronts based on eight quality indicators. In addition to four general multi-objective optimization quality indicators, the three aspects of fairness among clients and also uncertainty are reconfigured as quality indicators. These quality indicators are computed for a Pareto front. Results show that MOWOA performs better than others and makes requirement selection fairer. MOGWO works better than the rest when budget constraints are reduced.

选择一组要在下一个软件版本中实现的要求是一个称为NRP的NP-Hard问题。我们提出了多目标版本的灰狼优化器和鲸鱼优化算法来解决双目标NRP。我们使用了这两种算法和其他三种进化算法来解决来自四个数据集的NRP问题实例。成本分数比和轮盘赌用于满足NRP问题的约束。我们根据八个质量指标比较获得的帕累托前沿。除了四个通用的多目标优化质量指标外，客户之间的公平性和不确定性这三个方面也被重新配置为质量指标。这些质量指标是针对帕累托前沿计算的。结果表明，MOWOA的性能优于其他，并使需求选择更公平。