A novel particle swarm optimization with small world network and group decision information

Particle swarm optimization (PSO) is a daughter of artificial society and social learning. Hence, this paper excavates the ultimate source of PSO further, and then introduces the thinking of small world network and group decision information into it to obtain a new conceptual framework and algorithm variation for PSO, which is named PSO-WG. At the same time, the PSO-WG is discussed from the perspective of evolutionary computing to clarify the optimizing mechanism and improvement principles, which mainly includes the biological metaphor, implicit parallelism, operator mapping and feedback control analysis. Next, the computational model is proposed for achieve a self-contained optimization solution. Subsequently, a series of benchmark functions are tested and contrasted with the former representative algorithms to validate the feasibility and creditability of the new algorithm whose comprehensive performance is analyzed detailedly. Finally, the deficiency of PSO-WG and the working direction are pointed out clearly.