Research On Vehicle Routing Problem Based On Improved Particle Swarm Optimization Algorithm

Whether or local troops, logistics and distribution as the basis for logistical andeconomic development, its function and role was continuously strengthened. VehicleRouting Problem(VRP) is key to the overall logistics optimization. But also socialdevelopment, economic activity indispensable part. Reasonable vehicle routing cangreatly improve logistics and distribution of the economic benefits, savingdistribution costs.VRP is a typical NP-hard problem. A reasonable allocation of vehicles caneffectively improve transport efficiency, saving transportation costs, having greatsignificance to the economic development. In this paper, after inspection and analysisof the relevant literature and research results domestic and abroad, for the VRPconducted in-depth research, in this paper The main work is as follows:First, analysis the model of the VRP in distribution, according to currentresearch, given the definition of VRP, classification and algorithms for solving suchproblems.Then, in the study of domestic and foreign-based vehicle routing optimizationproblem-solving algorithms, Particle Swarm Optimization(PSO) algorithm easy tofall into local optimum characteristics and lack of local search capabilities defects, inthis paper, learn the idea of Bacterial Foraging(BF) algorithm, copy, extinction andmigration mechanism is introduced into the PSO algorithm. Proposed An improvedparticle swarm algorithm.Secondly, after the pilot test function to verify that the improved PSO in the sizeof certain problems can effectively improve the original particle swarm optimizationcapabilities and the ability to get rid of local optimum.Finally, the algorithm is applied to a specific vehicle routing optimization testinstances, be used to further verify its effectiveness and ability to solve practical problems. After five test experiments, the results show that the algorithm for solvingthe convergence speed and accuracy performance is relatively good, can be used as aneffective means to solve VPR.