Research Area: Metaheuristic Computing
In this work the well-known Particle Swarm Optimization (PSO) algorithm is applied to some Dynamic Optimization Problems (DOPs). The PSO algorithm is improved by simplification instead of introducing additional strategies into the algorithm as done by many other researchers in the aim of improving an algorithm. Several parameters (w, V max , V min and c 2) are being excluded from the conventional PSO. This algorithm is called Weightless Swarm Algorithm (WSA) as the prominent parameter, inertia weight w does not exist in this proposed algorithm. Interestingly, WSA still works effectively via swapping strategy found from countless trials and errors. We then incorporate the proven clustering technique from literature into the framework of the algorithm to solve the six dynamic problems in literature. From the series of tabulated results, we proved that WSA is competitive as compared to PSO. As only one parameter exists in WSA, it is feasible to carry out parameter sensitivity to find the optimal acceleration coefficient, c 1 for each problem set.
Keywords:
dynamic optimization
swapping
weightless swarm
Particle Swarm Optimization
acceleration coefficient
Author(s) Name: T. O. Ting, Ka Lok Man, Sheng-Uei Guan, Mohamed Nayel & Kaiyu Wan
Journal name:
Conferrence name: Network and Parallel Computing
Publisher name: Springer
DOI: 10.1007/978-3-642-35606-3_60
Volume Information: volume 7513
