Abstract: Integrating physical manufacturing resources and transforming to logic manufacturing resource was the main task for resource virtualization in cloud manufacturing, there was no feasible solution to this problem at present. Based on the mining of workflow log and definition of activities with resource configuration in workflow model, the service frequencies of physical manufacturing resources composition were calculated from two levels, business process and activity instance respectively. Then multi-granularity composition set was obtained by merging the two results. This approach could supply the reliable assurance for mapping from physical to virtual manufacturing resource. Finally an experiment was proposed to verify the effectiveness of this method.

Key words: cloud manufacturing, workflow, resource composition, resource virtualization, physical manufacturing resource

摘要： 把物理制造资源组合在一起并转化为逻辑制造资源是云制造中资源虚拟化的主要任务，目前尚无实用的资源组合方法，为此基于对工作流日志的挖掘，以及工作流模型中对活动及资源配置的定义，分别从过程实例和活动两个层面计算物理制造资源组合的使用频率，最后合并两个计算结果得到不同粒度的资源组合，以此形成的资源组合集可在资源虚拟化时为物理-虚拟资源的映射提供依据。最后通过实验验证了所提方法的有效性。

关键词: 云制造, 工作流, 资源组合, 资源虚拟化, 物理制造资源, 虚拟制造资源