Topology generation for knowledge-based fuzzy Petri nets

To implement fuzzy reasoning systems fuzzy Petri nets are very effective for the representation of expert knowledge, and then refined using evolutionary learning techniques. Most of these system, however, lack the ability to refine their network's topology and thus unable to add new rules to the reformulated rule base. To resolve this kind of problems, we propose a new algorithm to improve the initial domain theory of expert using a heuristic learning based on beam search. The proposed method generates a new node where the node is evaluated in the network. In the results, it can be extended to more general knowledge than the initial one. The proposed method is applied to a medical diagnostic system to verify its validity and effectiveness.