摘要
从海量候选服务组合计划中选择具有最优/满意服务质量(Qo S)的计划,即基于Qo S的Web服务组合,是面向服务计算的难题之一。为此,将该问题建模为多属性决策问题,同时针对传统多属性决策方法难以处理海量搜索空间的问题,提出一种结合折中比例法和遗传算法的新型智能进化算法(GACRM)。GACRM结合了折中比例法的方案排序优势和遗传算法高效的全局搜索能力,能够从海量搜索空间中快速找到全局近似最优解。实验结果表明,该算法不仅能够高效地产生与折中比例法接近的最优方案排序,且在解决大规模Web服务组合问题上具有良好的可伸缩性。
The problem of Quality of Service ( QoS)-based Web Service Composition ( QWSC), i. e. , selecting an optimal/satisfactory Service Composition Plan( SCP) from numerous candidate plans on the basis of QoS properties,is the most critical issue in the service-oriented computing. In this paper,the problem of QWSC is formulated as a Multi-Attribute Decision Making ( MADM ) representation. Furthermore, an intelligent evolutionary algorithm:Genetic Algorithm based Compromise Ratio Method ( GACRM ) is developed to solve the MADM problem. Combining with the advantage of Compromise Ratio Method( CRM) in terms of ranking alternatives,together with the superiority of Genetic Algorithm( GA) in terms of global search,GACRM is capable of finding an approximate optimal solution from a massive search space. Experimental result shows that GACRM is highly efficient and scalable for large-scale QWSC problems.
出处
《计算机工程》
CAS
CSCD
北大核心
2015年第6期33-37,42,共6页
Computer Engineering
基金
国家自然科学基金资助项目(71101103)
关键词
WEB服务组合
服务质量
多属性决策
折中比例法
遗传算法
Web service composition
Quality of Service ( QoS )
Multi-attribute Decision Making ( MADM )
Compromise Ratio Method (CRM)
Genetic Algorithm (GA)