Proposal of Parameter Setting Method on Independent-minded Particle Swarm Optimization
Haruna Matsushita · IEICE Proceeding Series · 2014
This study proposes a setting method of the important parameter which influences the optimization ability on an Independent-minded Particle Swarm Opti- mization (IPSO). The proposed parameter is a linear func- tion proportional to the simulation step. We confirm that although it is very simple and does not need additional pa- rameters, the proposed IPSO obtains better results than the standard PSO and the conventional IPSO, for the multi- modal functions. From these results, we do not need com- plicated settings of the parameters and can easily use the IPSO.