CONVERTING INFLUENCE NETS WITH TIMING INFORMATION TO A DISCRETE EVENT SYSTEM MODEL

Recent research has shown how to incorporate time in probabilistic modeling techniques called influence nets that are used to model complex political/military situations. By adding timing information to these models, which are static equilibrium models, they are converted to discrete event system models that can be represented as colored Petri nets. Executing these CP nets reveals the dynamic changes in the probability values of key events that are modeled as propositions in the influence net. This paper illustrates how the state space analysis capability of Design/CPN was used to verify the behavior of a generic CP net model generated from the influence net. First, the implications of incorporating time in an influence net are presented along with the type of behavior that the discrete event model should have. A procedure for interconnecting CP net modules to create the overall model is presented. Finally, the state space analysis capabilities of Design/CPN are used to verify the behavior of the model and reveal interesting properties of the dynamical model that are not intuitively obvious from the structure of the model.