A predicate/transition net model for artificial intelligence robot planning

This thesis consists of two parts. The first part is concerned with a place/transition net model for propositional logic rule-based systems. Inputs and outputs of a real-time control system are in general binary information. Inputs are 0 or 1 values from sensors, and outputs are 0 or 1 values to enable or disable actuators. For this application domain, propositional logic is a natural method of representing a problem specification. A propositional logic program can be easily modelled in a place/transition net. We propose the place/transition net fixpoint method for real-time control system design. The method has been implemented. Test results show that our algorithm is very efficient.;In addition, we propose a bidirectional means-ends analysis technique for our predicate/transition model for the planning problem. Conventional bidirectional search strategies are valid under the assumption that the goal is fully described. In planning problems, however, a goal is seldom fully described. The bidirectional search technique proposed in this dissertation can handle a partially described goal. We also provide admissible heuristic functions which can be used with heuristic search algorithms for performing pr/t net reachability tests. The iterative deepening A;The main problem considered in this thesis is the predicate/transition (pr/t) net method for artificial intelligence robot planning problems. We present an algorithm for automatic construction or a predicate/transition net model for a robot planning problem. Then by analyzing the net model, we can find a solution for the planning problem. Analysis of the net model is independent from the planning problem.