Research And Application Of Multi-objective Particle Swarm Optimization Based On Ring Topology

There exist many multi-objective optimization problems in the scientific research and engineering practice with the principle of maximizing the benefit and minimizing the cost.In recent years,the multi-objective optimization problem with high-dimensional variables has gradually become a hot-spot and difficulty in the field of multi-objective evolutionary algorithms.The existing evolutionary algorithm has such a significant decrease in its performance when dealing with high-dimensional variables that it is difficult to meet the practical need.Without the operation like the cross,mutation in genetic algorithm,Multi-objective Particle Swarm Optimization has such advantages as fast convergence,low parameter and simple calculation.Therefore,it has attracted great attention from scholars at home and abroad.But with the increase of decision variables on optimization problem,the algorithm will encounter with “dimension disaster.” To solve the above problems,this paper,based on the Cooperative Co-evolution Multi-Objective Particle Swarm Optimization(CCMOPSO),introduces the neighbor topology of the ring structure and the particle search pattern of the Gaussian-Cauchy distribution,and puts forward collaborative CCMOPSO algorithm with the cyclic structure.In addition,this paper studies the application of the algorithm in the grouping optimization problem of large-scale population,taking the group optimization problem in cooperative learning as an example.The main research results of this paper are summarized as follows:(1)In order to solve the problem that the multi-objective particle swarm algorithm is easy to fall into the local optimum,the algorithm is optimized from search method of the neighbor topology and the particle.By introducing the neighbor relation of the ring structure,the convergence rate of the algorithm is slowed down,which reduces the probability of “precocious” and falling into the local optimum.At the same time,the Gaussian and Cauchy distribution is used as the search pattern of the particle,replacing the position updating formula of original particle,increasing the probability that the particle group jumps out of the local optimum,and keeping the global and local search ability of the particle group balance.The simulation results based on the test function show that the convergence and diversity of the improved algorithm are improved at the same time.(2)As to the group optimization in cooperative learning,based on Lin's EPSO grouping optimization model,a collaborative CCMOPSO algorithm with the ring structure is proposed to solve the problem of the significantly decrease in the original algorithm performance and other issues when the number of groups increases sharply.Finally,by compared and analyzed with EPSO,exhaustive method and stochastic random grouping method,the results show that the cooperative CCMOPSO algorithm based on ring structure is less affected by the number of groups and the quality of the obtained scheme is better and the calculation time is reasonable,thus effectively solve the problem of group optimization on large number of cooperative learning.