求解服务等级感知服务组合问题的多目标遗传算法  

Multi-objective genetic optimization algorithm for SLA-aware service composition problem

作　　者：刘磊[1] 杨冬[1] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《吉林大学学报（工学版）》2015年第1期267-273,共7页Journal of Jilin University:Engineering and Technology Edition

基　　金：吉林省重点科技攻关项目(20130206052GX);国家自然科学基金项目(61300049);教育部高等学校博士学科点专项科研基金项目(20120061120059)

摘　　要：针对已有求解多等级服务部署问题的算法存在的不足,提出了一种求解该问题的多目标遗传算法(SMOGA),建立了求解该问题的多目标优化模型。通过将个体的支配强度和被支配强度结合到一起建立对个体的评价策略,并根据评价结果进行环境选择及生成个体的交叉概率。此外还设计了一种局部搜索策略并将其融入到变异策略中,以提高变异操作的有效性。最后对SMOGA算法进行了分析,并将该算法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SMOGA能够更加有效地解决该问题。The service deployment with multi-service level agreement is a basic problem to be solved for the service based software system. To overcome the shortcomings of the existing algorithms, a multi-objective model for this problem is built and a Simulation-based Multi-Objective Genetic Optimization Algorithm （SMOGA） is proposed. According to the characteristics of this problem, the genetic operations are redefined, including the introduction of individual domination strength into the environment selection, and the redefinition of crossover strategy and individual mutation strategy incorporated with local search. The performance of the proposed SMOGA was analyzed and compared with that of recently proposed algorithms E3-MOGA and NSGA-II for this problem on different scale test cases. Experiment results show that the proposed SMOGA can solve this problem more effectively.

关 键 词：计算机应用 多目标优化 遗传算法 服务等级 支配强度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]