摘要
Web服务软件工程的实用化挑战之一是QoS感知的选择、组合和稍后的绑定,表现为允许在运行时绑定一组领域Web服务构成面向服务的系统.这些领域Web服务在提供要求的功能同时,还满足一些非功能约束,例如总的费用或响应时间,并且使之最优化.对此作者提出了一种把Web服务看作为首类构件对象的关系查询基础结构,它通过各种Web服务操作调用评估查询.鉴于个性化和效率在这种评估中的重要作用,提出了一个基于聚合不同Web服务的多属性QoS参数的查询优化模型.该模型通过用户设定的全局约束和偏好、一个动态的等级方案以及多级匹配来调整QoS.等级提供了一个Web服务的行为评估,而多级匹配通过使用类似的和部分的答案对解决方案的空间进行扩展.进而给出了模型求解的遗传算法,并从适应度函数的静态惩罚、动态惩罚以及拉伸3个方面对优化性能进行了比较.文中最后介绍了一个从高端实现的服务查询引擎原型系统,用以展示该方法的适应性、可行性和有效性.
One of the most promising opportunities from a Web services engineering perspective is the QoS-aware selection, composition and late-binding. This allows you to dynamically assemble a collection of domain-specific QoS-aware Web services providing the required features into a composition services that can meet some non-functional constraints, and optimize criteria such as the overall cost or response time. This paper presents a query infrastructure that considers Web services as first class component objects. It evaluates queries through the invocations of different Web services operations. Because personality and efficiency play a central role in such evaluations, the paper proposes a query optimization model based on aggregating the multi-attribute QoS parameters of different Web services. The model adjusts QoS through global constraints and preferences set by the user, a dynamic rating scheme, and multilevel matching. The rating gives an assessment of Web services behaviors. Multilevel matching provides the expansion of the solution space by enabling similar and partial answers. The paper describes a genetic algorithm for solving the model, and compares the optimization performance of the genetic algorithm using various fitness functions varying in terms of static penalty, dynamic penalty, and stretching. The proposed approach has been applied to a service query engine system for SEIFCW, which consequently shows its applicability, feasibility and efficiency.
出处
《计算机学报》
EI
CSCD
北大核心
2009年第5期1014-1025,共12页
Chinese Journal of Computers
基金
国家"八六三"高技术研究发展计划项目基金(2002AA415280)
教育部博士点基金项目(20050359004)
教育部新世纪优秀人才计划项目(NCET-04-050562)
合肥工业大学科学研究发展基金项目(2007GDBJ012)资助~~
关键词
WEB服务
服务质量
服务选择
约束优化
遗传算法
Web services
quality of service （QoS）
service selection
constraint optimization
genetic algorithm
