A novel particle swarm optimiser with semi-active congregation and a nature-inspired supervising algorithm
Mohammad Kazem Akhlaghi, Alireza Fathi, Hamidreza Mohamadi Danyali, Mohamad Ali Khesali · International Journal of Computational Science and Engineering · 2014
Particle swarm optimiser (PSO) is one of the most accepted evolutionary computation algorithms that have attracted incremental interest from scientists and engineers. The paper presents a new particle swarm optimisation technique combined with a supervising algorithm for controlling the particle velocity. The proposed variant, which is called PSO with semi-active congregation (PSOSAC), is a mathematical model of animals’ congregation behaviour in nature. Observing the response of a school of fish in facing an attack inspired the authors to develop a supervising algorithm for adjusting the particles’ velocity. Also the proposed algorithm allows narrowing the search space. The authors evaluate the proposed algorithm with ten different benchmarks, and compare it to other well-known PSO variants. The results are very promising, demonstrating superior results when compared to other techniques, and a good convergence rate.