Dynamic program analysis based on partitioning and summarization
基于分段执行和摘要的动态程序分析*

In a modular system design, the models of computation appropriate for the components usually vary. Those models of computation include finite state machine, Petri net, discrete events and event relationship graph. To provide flexibility for designers and to improve productivity, it is necessary to enable the use of different models of computations in the components, and to apply the theories of composing models of computation to construct complete models for simulation and automatic system generation. As an application, by hierarchically composing discrete events and event relationship graphs, dynamic systems that were easy to extend, modify and maintain could be created. The theories could also be applied to other models of computation, and thus exhibited their benefits in compositional designs.