Complex nonlinear system identification based on cellular particle swarm optimization

Abstract

Nonlinear system identification is one of the most important topics of modern identification. A novel approach for complex nonlinear system identification is proposed based on cellular automata particle swarm optimization (CAPSO) algorithm in this paper. The problems of nonlinear system identification are converted to nonlinear optimization problems in continual space, and then the PSO algorithm is used to search the parameter concurrently and efficiently to find the optimal estimation of the system parameters. In order to enhance the performance of the PSO identification, an improved PSO based on cellular automata is proposed by combining cellular automata (CA) with PSO. In the proposed CAPSO, each particle of particle swarm is considered as cellular automata, and distributes in the two-dimensional grid, and the state update of each cell is not only related to its own state and the state of neighbors, but also considers the state of the optimal cell. If the state is too close with the optimal cell, then re-update the cell state. The simulation results show the effectiveness and the feasibility of the proposed method.