多目标人工蜂群算法在服务组合优化中的应用

应用广泛的聚集函数法可将多目标问题转换为单目标问题,但函数设计困难,通用性不强,且优化结果不能直观反映各目标的优化情况。提出了一个基于Pareto占优的多目标人工蜂群算法,改进了邻域搜索策略,给出一个对比实验。实验结果表明,改进算法在个体多样性与对Pareto最优边界的趋近程度方面均有优势。

Multi-objective artificial bee colony algorithm applied in QoS-awareservice composition optimization

The widely used aggregate function algorithms can transform multi-objective problems into single-objective problems, but it's difficult to be designed, the versatility is weak, and the results can't reflect the optimization of each of the target. This paper proposed a multi-objective artificial colony algorithm based on Pareto, and improved the neighborhood search method. Finally, it gave an experiment. The results show that the improved algorithm has an advantage in the diversity of individuals in group and the approachability with the Pareto frontier.