Enhancing the NichePSO

The NichePSO was developed as one of the first particle swarm optimization (PSO) approaches to locate multiple solutions to continuous optimization problems. The NichePSO forms subswarms from a main swarm, where each subswarm represents a single niche (or solution). Mechanisms are employed to merge subswarms if they converge to the same solution, and also to absorb any particle within a subswarm if that particle enters the area covered by the subswarm. The NichePSO has shown very good performance in locating a good number of solutions to multimodal problems. However, it was found that the current subswarm merging and particle absorption strategies are premature, and limits exploration in the main swarm. This paper proposes a number of different merging and absorption strategies, and shows that fine tuning of these processes improves the performance of NichePSO on lower dimensional problems.