A New MOPSO Based on Pairing Selection and Adaptive Strategy

Abstract

In order to improve the performance of particle swarm optimization, aim at the poor convergence rate and the poor local optimum search capabilities, proposing an improved multi-objective particle swarm optimization. The algorithm is based on the information transmission mechanism between particle swarm, uses SPEA2 environmental selection and pair selection strategy in algorithm to make the population of particles quickly converge to Pareto optimal boundary and uses adaptive principle to change the calculation method of the speed weight to enhance the algorithm's global search capability. Through the simulation experiments of classic test functions and the application of robot path planning, the results show that the improved algorithms make the algorithm not only makes it easier to jump out of the local algorithm but also makes the convergence speed of algorithm and the convergence speed of particle populations have been greatly improved, also makes the robot path planning algorithm can more quickly find the optimal road king.