一种求解SLA等级感知服务组合问题的多目标离散粒子群优化算法

A Multi-Objective Discrete Particle Swarm Optimization Algorithm for SLA-Aware Service Composition Problem

针对SLA等级感知服务组合问题,本文提出了一种求解该问题的多目标离散粒子群算法(MDPSO),建立了多目标粒子群算法优化模型.根据该问题的特征,对粒子更新策略进行重新设计;并且提出粒子变异策略以抑制群体的早熟收敛增强群体的全局搜索能力.另外,提出了一种基于约束支配关系的局部搜索策略并将其结合到MDPSO算法,形成算法MDPSO+.最后对MDPSO算法的参数设值进行了分析,并将算法MDPSO、MDPSO+与最近提出的求解该问题的E3-MOGA算法及NSGA-II算法在不同规模的测试用例上进行了实验对比,结果表明算法MDPSO+能够更加有效的解决该问题.

For SLA-aw are service composition problem（ SSC）,a multi-objective discrete particle sw arm optimization algorithm（ M DPSO） is proposed in this paper and an optimization model for this algorithm is also built.According to the character of this SSC problem,a particle updating strategy is redesigned by introducing crossover operator.A particle mutation strategy is proposed to increase the sw arm diversity and restrain particle sw arm ＇s premature convergence.In addition,algorithm M DPSO ＋ is formed by incorporating a local search strategy based on constraint-domination into the algorithm M DPSO.At last,some parameters in algorithm M DPSO are analyzed and set w ith relative proper values,and then the algorithm M DPSO and the algorithm M DPSO ＋ are compared w ith the recently proposed algorithm E3-M OGA and NSGA-II on different-scale cases; the results show that algorithm M DPSO ＋ can solve the SSC problem more effectively.