基于MapReduce改进蚁群算法的Web服务组合优化  

Web service composition optimization based on improved ant colony algorithm with MapReduce

作　　者：颉斌[1] 杨扬[1] 王洁莹[1] 

机构地区：[1]北京科技大学计算机通信工程学院,北京100083

出　　处：《微型机与应用》2016年第8期61-64,共4页Microcomputer & Its Applications

基　　金：国家自然科学基金(61272508)

摘　　要：对于Web服务组合优化的问题,蚁群算法的求解主要是串行进行,收敛时间长,容易收敛于非最优解。在云计算环境中,将蚁群算法并行化,可对Web服务组合优化问题进行分布式并行求解。根据多目标优化模型给出基于多信息素的蚁群算法,使用MapReduce并行编程框架对蚁群算法中最耗时的部分——蚂蚁独立求解的过程并行化,给出了使用MapReduce改进的基于多信息素的蚁群优化算法,有效地对Web服务组合进行全局优化,弥补传统的蚁群算法求解过程的缺点。As to the problem of Web service composition optimization,ant colony optimization（ ACO） is usually solved serially. However,it has long convergence time and is likely to converge to a non-optimal solution. In Cloud computing environment,with parallelized ACO Web service composition optimization can be solved in parallel. In this paper,it is abstracted to a multi-objective optimization problem,and a mathematical model is built. According to the multi-objective optimization model,a multi-pheromone algorithm based on ACO is given. MapReduce parallel programming framework is adopted to parallelize the most time-consuming part of ACO-independent ant solving. Finally,the improved multi-pheromone with MapReduce based on ACO is given to effectively perform global optimization to Web service composition,which overcomes the shortcomings of traditional ACO.

关 键 词：服务组合 服务组合优化 蚁群算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]