A Modified Particle Swarm Optimization with a Smart Particle for Inverse Problems in Electromagnetic Devices
Rehan Ali Khan,Shiyou Yang,Shah Fahad,Shafi Ullah Khan,Kalimullah
DOI: https://doi.org/10.1109/access.2021.3095403
IF: 3.9
2021-01-01
IEEE Access
Abstract:Particle swarm optimization (PSO) is a swarm intelligence-based metaheuristic algorithm inspired by the natural behavior of birds flocking or fish schooling. The PSO’s main advantages are its ease of implementation and a small number of fine-tuning parameters. However, the major drawbacks of an existing PSO are its premature convergence and the lack of a balance of exploration and exploitation searches in the search space. To address the aforementioned problems, a new concept known as a smart particle swarm optimization (SPSO) process is introduced and implemented. The smart particle that leads the swarm in the proposed concept has eidetic memory behavior. The smart particle mainly works under the principles of a convergence factor (CF) technique, which integrates the memorization of particles positions, comparison, and leader declaration for the best optimal solution. Additionally, CF uses a particle position vector instead of a particle fitness or mutation to increase the exploration capability in the search space. The TEAM Workshop Problem 22, a super conducting magnetic energy storage (SMES) system; and some well-known benchmark optimization test functions are numerically solved to verify the efficacy of the proposed SPSO. The SPSO finds a better optimal solution than the other tested algorithms, particularly in the initial computational evaluation of the generation according to numerical experiments and case study analysis.