中文摘要：

为了从面向对象的遗留系统中识别出高质量的服务,提出一种半自动化的遗留系统的服务识别方法.通过分析构件识别方法和服务识别方法,以业务流程模型和统一建模语言类图作为服务识别方法的输入,将服务识别问题转换为业务流程图的划分问题.为了得到一组图划分的最优解即最优服务集,又将图划分看作一个多目标优化问题.分别定义了服务集粒度、耦合度和内聚度的定量化关系,建立了多目标优化的服务识别模型.通过线性加权,给出了基于粒子群优化算法的服务识别算法.以某工程遗留系统为例,通过对比实验验证了方法的有效性.

英文摘要：

To identify the high quality services from object oriented legacy systems,a semi-automated service identification method of legacy system （SIOL） was proposed.Through analyzing the Component Identification Method （CIM） and Service Identification Method （SIM）,the problem of service identification was transformed into partitioning problem of business process graph by taking business process model and Unified Modeling Language （UML）diagrams as input.To get the optimal solution of graph partitioning,the graph partitioning was regarded as multiobject optimization problem.The quantitative relationship of service＇s granularity,coupling and cohesion was defined respectively,and service identification model of multi-object optimization was built.By linear weighted,a service identification algorithm for legacy system based on particle swarm optimization algorithm was given.The proposed method was used for designing web services in a legacy system,and the effectiveness was verified by comparison experiment.