Open Access

ARTICLE

Employing a Diversity Control Approach to Optimize Self-Organizing Particle Swarm Optimization Algorithms

Sung-Jung Hsiao¹, Wen-Tsai Sung^2,*

1 Department of Information Technology, Takming University of Science and Technology, Taipei City, 11451, Taiwan
2 Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung City, 411030, Taiwan

* Corresponding Author: Wen-Tsai Sung. Email:

Computers, Materials & Continua 2025, 82(3), 3891-3905. https://doi.org/10.32604/cmc.2025.060056

Received 22 October 2024; Accepted 21 January 2025; Issue published 06 March 2025

Abstract

For optimization algorithms, the most important consideration is their global optimization performance. Our research is conducted with the hope that the algorithm can robustly find the optimal solution to the target problem at a lower computational cost or faster speed. For stochastic optimization algorithms based on population search methods, the search speed and solution quality are always contradictory. Suppose that the random range of the group search is larger; in that case, the probability of the algorithm converging to the global optimal solution is also greater, but the search speed will inevitably slow. The smaller the random range of the group search is, the faster the search speed will be, but the algorithm will easily fall into local optima. Therefore, our method is intended to utilize heuristic strategies to guide the search direction and extract as much effective information as possible from the search process to guide an optimized search. This method is not only conducive to global search, but also avoids excessive randomness, thereby improving search efficiency. To effectively avoid premature convergence problems, the diversity of the group must be monitored and regulated. In fact, in natural bird flocking systems, the distribution density and diversity of groups are often key factors affecting individual behavior. For example, flying birds can adjust their speed in time to avoid collisions based on the crowding level of the group, while foraging birds will judge the possibility of sharing food based on the density of the group and choose to speed up or escape. The aim of this work was to verify that the proposed optimization method is effective. We compared and analyzed the performances of five algorithms, namely, self-organized particle swarm optimization (PSO)-diversity controlled inertia weight (SOPSO-DCIW), self-organized PSO-diversity controlled acceleration coefficient (SOPSO-DCAC), standard PSO (SPSO), the PSO algorithm with a linear decreasing inertia weight (SPSO-LDIW), and the modified PSO algorithm with a time-varying acceleration constant (MPSO-TVAC).

---

Keywords

---