The Improvement And Application Of Multi-objective Particle Swarm Algorithm

Particle swarm algorithm is a new heuristic global optimization technology based on group optimization, its basic idea comes from the simulation of birds foraging process in nature. Particle swarm algorithm was attracted wide attention owing to the simple model, the advantages of fast convergence speed and easy implementation after it is proposed. At the same time, the applications of PSO are also more widely, including optimization function, industrial systems optimization and control and the applications fields of other evolutionary algorithm and so on. Various studies on its application and advantages have emerged what promoted the study of PSO develop effectively. As robots become the most representative of the highest technical areas of the strategic objectives, robot path planning as its navigation technology is an important research topic has aroused extensive research. With a variety of evolutionary algorithms applied research in robot path planning, PSO algorithm is gradually being applied to research in this field.Aiming at some shortcomings of standard MOPSO, such as the great randomness of the best individual historical value position and the global optimal value position selection for PSO, and the poor global search ability and poor local search ability of the population and so on what is the reason for the improved algorithm. The main research work of this paper including the following aspects:According to the fitness value idea in the single objective particle swarm algorithm, used the fitness value to evaluate the merits of the idea of particle, using the calculation method of SPEA2to give the fitness value for each particle what provide an evaluation of multi-objective particle swarm optimization in the multi-objective particle swarm algorithm.According SPEA2environmental selection and pairing selection strategy to solve the random problem of the best individual historical value position and the global optimal value position selection.Using the adaptive method to change the calculation method of the speed of the weight what enhance the activity of populations of particles in local optimum and what lead the algorithm has strong global search ability and local search capabilities.According to the idea of chromosome in GA and the idea of bisector to specification the particle position, and establish the environmental models and mathematical models of the robot path planning.Using the validation of the algorithm to the path planning of robot, and using the robot path planning results in turn validate the feasibility and validity of the improved algorithm.