Piecewise System Modeling Algorithm Based on Simplified Fuzzy Inference Reasoning and Particle Swarm Optimization

This paper proposes a piecewise modeling method based on fuzzy if-then rules using particle swarm optimization. The piecewise model has the shape of a rectangular partition of the state-space; the model can be represented as a fuzzy if-then rule with singleton consequents. The vertex values of the rectangular regions are determined using particle swarm optimization because the optimal solution is a nonlinear programming problem. In order to determine optimal vertex values of the piecewise regions with minimal modeling errors, this paper proposes a learning algorithm based on simplified fuzzy inference reasoning and particle swarm optimization methods. Some numerical simulation results show the effectiveness of the learning algorithm.