A Memory-Based Particle Swarm Optimization for Parameter Identification of Lorenz Chaotic System

A novel modified version of particle swarm optimization (PSO) is introduced in this paper to estimate the parameters of the chaotic Lorenz system. The parameters estimation of the Lorenz system is modeled as a multidimensional problem and solved by the proposed algorithm, a memory-based particle swarm optimization (MbPSO) algorithm. In MbPSO, two new terms are added to the standard PSO to vary the population direction and enhance search capability. Firstly, the impact of parameter configuration on MbPSO is studied. After that, the parameter estimation problem is solved. The performance of the proposed MbPSO is compared with other meta-heuristic algorithms in terms of parameter accuracy and convergence speed. According to the results, linking the memory of each particle to the memory of other particles has a very significant effect on the proposed algorithm compared to the original PSO. Briefly, the MbPSO algorithm is a successful and powerful optimization algorithm for parameter estimation of chaotic systems with accurate performance.