基于蚁群算法的面向服务软件的部署优化方法  

Deployment Optimization of Service-Oriented Software Based on Ant Colony Algorithm

作　　者：李琳[1,2] 应时[1,2] 赵翀[1,2] 董波[1,2] 

机构地区：[1]武汉大学软件工程国家重点实验室,湖北武汉430072 [2]武汉大学计算机学院,湖北武汉430072

出　　处：《电子学报》2016年第1期123-129,共7页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61373038;No.61070012);国家863高技术研究发展计划(No.2012AA011204-01)

摘　　要：面向服务软件的部署优化问题是典型的NP难题.本文构建了基于性能改善的软件部署优化模型,设计了一种蚁群优化算法ACO-DO进行近似最优解的快速求解.该算法通过设计基于部署优化问题的启发式、改进部署方案的构建顺序、增加局部搜索过程实现蚁群算法求解效率的提升.通过不同规模的实例实验,验证了ACO-DO算法能够取得比现有的混合整数线性规划算法、蚁群算法和遗传算法更好的性能.The deployment optimization of service-oriented software is well known to be NP hard. In this paper,a software deployment optimization model is built for improving the performance of service-oriented software,and an Ant Colony Algorithm for Deployment Optimization（ ACO-DO） is designed to solve it so that the near-optimal solutions can be obtained quickly. The algorithm improves ant colony algorithm by designing a heuristic based on the considered problem,optimizing the orders of constructing deployment solutions and adding a local search procedure. A series of instances with different sizes are tested and analyzed. The experimental results showthat the designed ACO-DO algorithm performs better than the existing Mixed Integer Linear Programming,ant colony and genetic algorithms.

关 键 词：面向服务的软件 部署优化 蚁群算法 性能 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]