Abstract:

The multi-neighborhood improved particle swarm optimization algorithm (MNI-PSO) is proposed for the purpose of improving the capability of the standard particle swarm optimization (PSO). The MNI-PSO contains a simple neighborhood topology and sets an improved update scheme of the velocity inertial weight. The velocity and searching area restriction algorithms are also proposed for the MNI-PSO. The optimization results of the classical testing problems show that the MNI-PSO has performed a great capability for the complex optimization problems. Finally, the solution to multi-objective optimization problems (MOOP) using MNI-PSO is proposed. The classical 5-D optimization problem and Golinski’s speed reducer problem are optimized. The results show that the MNI-PSO’s performance is better than some other popular algorithms.