Predicate/Transition Net Based Model for Reasoning about Actions

A novel representation framework has presented in this paper. It bases on a special type of Petri nets called Predicate/Transition (Pr/T) nets for describing reasoning about actions for robots. This framework blends together the concept of the state of the world, actions and the inner state of the agent. It is capable of describing reasoning about actions for robots which inhabit a dynamic and uncertain environment. This Pr/T net formalism can represent the Frame Problem graphically and solve it effectively. Analysis techniques of Pr/T net are utilized to generate action sequences to achieve the goal. Additionally, our net model is graphical and easier to understand than other representations.