Models-Based Optimization Methods for the Specification of Fuzzy Inference Systems in Discrete EVent Simulation
Résumé
In this paper, we present our work in the field of computational intelligence and discrete event systems. Knowledge representation and the inclusion of imperfect knowledge is a key step in an effort to optimally incorporate artificial intelligence methods in a modeling and simulation framework. Fuzzy Inference Systems (FIS) are one of the most used applications of Fuzzy Logic and Fuzzy Sets Theory. They have the advantage of relying on the properties of Fuzzy Logic to represent imperfect information so gradually, and manipulate them from a linguistic description. This flexibility of representation is more significant for the study of complex systems. We wish to extend the Discrete EVent system Specification (DEVS) formalism to represent FIS and we propose a modular and generic approach (DEVFIS) to integrate in this new extension various optimization methods.