Abstract: The life cycle of cloud manufacturing involved the model-based activities. Under the environment of Internet oriented cloud, a large number of black box, heterogeneous models were accumulated, and the traditional retrieval and match methods could not satisfy the demand of massive model management. Thus the model composition should be taken into account in the process and the correctness of the time dynamic characteristics after composition should be ensured. Based on the composition approach of top-level modeling, semantic descriptions to each event of interactive scenarios among models were added, and existing redundant and mutually exclusive in the composition process were eliminated by planning method. Thus the combination of dynamic behavior was realized. The application example showed that the characteristics of highly scalable and on-demand aggregation could be achieved in the model resources management of cloud manufacturing by introducing models automatic composition technology.

Key words: cloud manufacturing, automatic composition, planning, models

摘要： 云制造全生命周期活动中都涉及到基于模型的活动。在面向互联网的云环境下会积累大量的黑盒异构模型，传统的检索匹配方法已不能满足模型管理的需求，需要在海量模型的管理过程中，同时考虑模型的组合问题并确保模型组合以后时间动态特性的正确性。为此，在基于顶层建模理论的模型组合方法的基础之上，通过为模型交互情景的各个事件添加语义描述，并利用规划的方法排除组合过程中存在的冗余和互斥，实现模型动态行为的组合。应用示例表明，通过模型自动组合技术的引入，可以实现云制造在模型资源管理方面的高可扩展和按需组合。

关键词: 云制造, 自动组合, 规划, 模型