Scalability Analysis of Competing Niching Paradigms in Swarm Intelligence: Sequential vs Parallel

Parallel niching has received considerable emphasis over sequential niching due to its ability to discover multiple basins of convergence with a lower investment in search agents, thus saving computational resources. The current research study argues that sequential niching is still viable owing to its scalability over time and good exploitation capability. Two swarm intelligence algorithms are compared to assess the scalability of both approaches: sequential niching particle swarm optimization (PSO) is compared against three representative parallel niching techniques NichePSO,kPSO, and Euclidean Distance Hierarchical Clustering PSO (EDHC) with a particular focus on the peak detection ratio of both paradigms.