摘要
针对服务组合中冗余服务池开销较大、服务响应时间较长的问题,提出一种多目标优化策略,在满足可靠性约束的前提下,优化服务池配置、降低冗余服务池开销和缩短服务响应时间。首先分析了在两种不同容错策略下,满足服务组合特性的服务池多个优化目标;然后采用基因多目标优化算法进行优化,全局采用非支配排序策略,局部采用随机漫步搜索策略。仿真实验通过与现有的研究方法进行对比,优化效果较好,表明该方法可用于优化冗余服务池配置。
Aimed at the problem that redundant service pool in service composition brings large cost and long response time,this paper presented a multi-objective optimization strategy to reduction the redundant cost and response time under certain reliability.At first,it analyzed the optimization objective of service pool used two kinds of fault tolerant strategies.Then,it used a memetic algorithm based on global nondominated sorting strategy and local random-walk method to solve the optimization problem.At last,simulation shows the strategy is validity and better than the other multi-objective evolutionary algorithms.
出处
《计算机应用研究》
CSCD
北大核心
2012年第10期3791-3793,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(90818028)
河南省重点科技攻关项目(102102210020)
关键词
可靠性
冗余服务池
服务响应时间
冗余开销
reliability
redundant service pool
service response time
redundant cost
