摘要
针对组合服务QoS优化中单目标优化建模需要精确定义QoS权重和多目标优化建模返回的Pareto最优解集数目过多的问题,借鉴多目标遗传算法(MOGA)和多属性决策领域的理想点法(TOPSIS),提出了一个能够快速返回Top-k最优组合服务、适用于较大规模服务组合问题求解的使用了TOPSIS的多目标遗传算法——T_MOGA。该算法在MOGA的方案评估阶段引入理想点法对组合服务进行评估,并通过选择、交叉、变异等操作对种群迭代以获取QoS更好的组合服务方案。T_MOGA有效简化了MOGA的非支配排序过程,减少了算法运行时间,并只返回Pareto最优解集Top-k最优组合服务,方便用户选择。通过实验,从算法的运行时间及得到集合的质量两个方面验证了T_MOGA的有效性。
By reference to the multi-objective genetic algorithm（MOGA） and the technique for order preference by simi- larity to ideal solution（TOPSIS）in multi-attribute decision making, the T_MOGA, a MOGA using the TOPSIS, was presented for optimization of composition service, without the difficulties that the modeling of single-objective optimi- zation needs accurate QoS weight defining and the number of Pareto optimal sets is too large when modeling of multi-objective optimazation. This T_MOGA can efficiently find the Top-k optimal solutions from the view of TOPSIS evaluation. It introduces the TOPSIS to evaluate the fitness of the service composition plans, and adopts the opera- tions of binary tournament selection, one-point crossover and bit-mutation to evolve plans with better fitness. The T MOGA simplifies the non-dominated sorting process of the MOGA, which reduces the running time of the algorithm, and it only returns Top-k optimal plans, which is convenient for users to select. The effectiveness of the algorithm in efficiency and scalability was verified by experiment.
出处
《高技术通讯》
CAS
CSCD
北大核心
2014年第2期131-137,共7页
Chinese High Technology Letters
基金
863计划(2011AA120302)资助项目
关键词
服务组合优化
TOP-K
理想点法(TOPSIS)
多目标遗传算法(MOGA)
服务质量(QoS)
service composition optimization, Top-k, technique for order preference by similarity to ideal solu-tion （ TOPSIS）, multi-objective genetic algorithm （ MOGA ）, quality of service （QoS）
