Inference Optimization Approach in Fuzzy Inference Systems

Abstract

This paper tries on the representation of Mamdani kind fuzzy inference systems by means of fuzzy association equations.The methodology is described to represent the control system modeled on fuzzy rules, in a system outlined by equations and inequations. Once represented the system here with the described methodology, these same equations are used to validate the system fuzzy rules. Later a linear programming software is used to solve the equations system and this way to carry out the fuzzy inference. The objective function of the equations system is also optimized through a learning algorithm based on upward search. This approach is demonstrated with an application example.