Abstract: To improve the collaboration of resource services among organizations,oriented to collaborative task systems,an approach for resource service sequence mining was proposed from workflow model and runtime data two levels.According to the task-related dependences between resource services in workflow model,an algorithm was proposed to resolve resource service sets,and an initial set of resource service sequences named resource service chain was obtained.To improve reusability of resource-service sequence,the initial resource service chain was taken as cluster center to extend all the resource service chains,and each new extended chain was operated to solve the similarities.Thus the modified cluster center was obtained.The effectiveness of proposed method was proved with experiment,and the analysis results showed that it could improve the selection efficiency,availability and utilization of resource services.

Key words: collaborative task, resource-service sequence, workflow, mining

摘要： 为提高组织之间资源服务的协同性,面向协同任务系统,从工作流模型和运行数据两个层面提出一种资源服务时序挖掘方法。通过分析工作流模型,依据资源服务间存在的任务相关性,求解依赖关系较强的资源服务集,并从中获取初步的资源服务序列,即资源服务链;为提高资源服务时序的重用性,以初始资源服务链为聚类中心,对其进行扩展后,求解运行数据中每个新到达扩展链的相似度,最终得到修正后的聚类中心。通过实验和分析验证了方法的可行性,表明该方法对提高资源的选取效率以及资源的可用性和利用率有积极的意义。

关键词: 协同任务, 资源服务序列, 工作流, 挖掘