中文摘要：

服务计算相关技术标准的持续完善和不断成熟推动了基于Web服务重用的分布式应用系统开发方式的迅速普及.而随着服务数量的爆炸性增长,网络上存在着大量功能相似、非功能特性各异的服务,如何在功能相当的服务集中选择质量较优的服务成为一个亟待解决的问题.传统的基于服务质量的服务选择方法,无论是局部最优或是全局最优策略,均面向服务库中的所有服务进行选择,选择效率受服务数量影响较大,因此不适用于基于大规模服务库的服务选择.文中引入数据库查询中的skyline方法,利用skyline中的支配关系,在选择过程中仅考虑skyline之上的服务,从而大大缩小了服务选择的范围,提高了服务选择的效率.同时针对动态Web服务环境,提出一种动态环境下的skyline服务维护算法,并通过一系列仿真实验证明了所提算法的高效性及良好的可扩展性.

英文摘要：

With the blossom of Web services,there are many function-equivalent services with different QoS（quality of service）.It has become a challenge to select services with high-quality from a set of function-equivalent services.Traditional approaches to service selection,with either partial or global optimizing strategy,process selection on all candidate services.These approaches are not suitable for selection oriented to large-scale services,as the efficiency is drastically limited by the number of services.This paper introduces the skyline approaches to improve the efficiency of selection by using the dominance relationship of skyline to prune services.It also proposes a novel skyline maintain algorithm which is suitable for dynamic service environment.An extensive performance study using synthetic data is reported to verify its efficiency.