Research And Application Of Multi-objective Particle Swarm Optimization Algorithm

Complex problems in real life often contain multiple goals.These goals are interrelated and conflicting.For such problems,many scholars try to use swarm intelligence algorithms to solve them.Particle swarm optimization is favored by researchers because of its powerful search ability and convergence ability.Therefore,many scholars try to use multi-objective particle swarm optimization to solve multi-objective optimization problems.However,the rapid convergence of particle swarm optimization makes its particles easy to fall into local optimum,and the Pareto front of multi-objective particle swarm optimization also has problems such as uneven distribution and poor diversity.Therefore,in view of these problems,this paper improves the multi-objective particle swarm optimization algorithm,and applies the improved multi-objective particle swarm optimization algorithm to feature selection and logistics path optimization of commodity customs clearance.Improve data classification performance and solve the problem of logistics path selection for commodity customs clearance.The main work of this paper is as follows:(1)In the multi-objective optimization problem,due to the defects of the traditional crowding distance,the uneven distribution of Pareto fronts is caused.In order to solve this problem,this paper proposes a new concept of crowding distance and improves the calculation formula of crowding distance.The new crowding distance no longer calculates the crowding distance based on the two closest particles to the particle,but calculates the crowding distance based on the first particle and the last particle,which not only improves the diversity of the solution but also makes the solution more diverse.distribution is more even.In addition,this paper also proposes a new mutation mechanism.Based on the traditional Cauchy mutation,a parameter that can adaptively adjust the variable asynchronous length is set according to the average moving speed of the population.The variable asynchronous length improves the search ability of particles in the population,and in the later iteration of the population,the population is disturbed with a smaller variable asynchronous length to speed up the convergence of the algorithm.(2)Apply the improved multi-objective particle swarm algorithm to the feature selection problem,and test it on 8 test data sets respectively,and compare the obtained number of features and classification error rate with the other three algorithms,which proves that this paper The proposed algorithm is aimed at the feature selection problem,and can obtain lower classification errors on the basis of reducing the number of feature extractions.(3)The improved multi-objective particle swarm algorithm is applied to the optimization of the logistics path of commodity customs clearance in each waybill at the import and export port,so as to optimize the customs clearance route.Experiments show that the algorithm proposed in this paper can reduce the total cost of customs clearance while reducing the time cost.