Parameter Optimization Strategy of Fuzzy Petri Net Utilizing Hybrid GA-SFLA Algorithm

Abstract

Fuzzy Petri net (FPN) is a powerful tool to model and analyze the knowledge-based systems (KBSs) or expert systems (ESs). The accuracy of the reasoning result is a bottleneck to hinder the further development of FPN because of lacking self-learning capability. To overcome this issue, a hybrid GA-SFLA algorithm is proposed in this paper to improve the precision of each parameter of a given FPN model. The proposed algorithm combines the advantages both of GA and SFLA and includes three phases, which are generating chromosome by encoding the multi-dimensional solution which reflects all initial frogs, gaining a better individual as well as seeking the optimal solution by executing the local search and global search operations of SFLA. Finally, an FPN model is used to test the feasibility of the proposed algorithm. Simulation results reveal that all parameters of the given FPN model have the higher precision by implementing the GA-SFLA than that of implementing GA and SFLA, respectively.