Abstract: For the question that the executable tools of Model Driven Architecture(MDA) such as xUML, OCL, ASL are too emphasis on the semantic integrity to observe and understand for designer, a human-computer interaction supported meta model Visual Action Model(VAM) is proposed by reconstructing UML model element in MOF2, and an executable engine of VAM is developed, which makes the process of the model code and the process of actual implementation code very similar, so as to enhance availability of requirement validation and test in model-driven development.

Key words: Model Driven Architecture(MDA), model element, Visual Action Model(VAM)

摘要： 针对xUML, OCL, ASL等模型驱动的执行工具过于侧重语义完整，不够直观，使得设计人员难以理解的问题，在MOF第二层基础上改造UML元模型元素，建立一种支持人机交互的元模型可视化动作模型(VAM)，在此基础上实现一个VAM执行引擎。模型的执行过程和实际代码的执行过程相似，提高了模型驱动开发在需求验证和测试中的可用性。

关键词: 模型驱动架构, 模型元素, 可视化动作模型