Particle swarm optimization (PSO) has been applied with success to many numerical and combinatorial optimization problems in recent years. However, a great deal of work remains to be done to improve the particle swarm performance. The purpose of this paper is to present a new adaptive PSO approach to overcome convergence drawbacks. Thus, the updating of the particle position rule and the introduction of new acceleration parameter augment the performance of the proposed model developed in this perspective.
In the studied picture, each particle defined in a multidimensional search space is represented by a vector of three adaptive parameters representing, respectively, the adaptive cognitive factor, the adaptive social factor, and the bi‐acceleration factor. Therefore, to updating its position rule, the authors add a gaussian noise to each updated velocity in order to increase the diversity in the population swarm.
The simulation experiments uses the CEC, 2005 functions benchmark. The achieved results show that the proposed model improves the existing performance of other algorithms compared to the same benchmark.
The proposed algorithm improves the performance of the PSO based on the self‐adaptation strategy. Thus, it can actually resolve hard functions which introduces noisy and shifted functions.
