Study And Application Of Improved Genetic Algorithm In Packages Distribution

With the fast development of the logistics industry, the Vehicle Routing Problem (VRP) is subjected to the concern by researcher and governor insreasingly. In the field of packages distribution, with the development of just in time, distribution cost enlarges gradually in proportion in the logistics cost, so the optimization of vehicle routes and cutting down distribution cost, having already become the main measure that controls the logistics cost.This papre aimed at studying Genetic Algorithm, proposed a kind of improved Genetic Algorithm sutible to VRP, and this way can be better to resolve VRP, provided the decision support to packages distribution, according to that purpose, the main done in this paper as follow :(1) The domestic and international concerning VRP was researched, established to take Genetic Algorithm as basic optimization way. Studied the rasearch way of VRP.(2) Genetic Algorithm was rsearched systematically, analyzed Genetic Algorithm basic carring out process, enphasized to rsearch the differentination between parthenogenesis algorithm and basic Genetic Algorithm, discovered parthenogenesis algorithm more suitable to solve VRP.(3) Aimed at the characteristic of VRP, combined to application exampie, from the angle of algorithms design to commence, analyzed the influence of optimization effect caused by algorithms, pointed out the intrinsic defect of standard Genetic Algorithm to VRP , then provided theoritical basis for the algorithms improvement design.(4) Aimed at aboved-mentioned defect, carried on improvement design of standard Genetic Algorithm.The main method was using the ray scan to creat initial population, combined to the advantange of parthenogenesis algorithm, designed converse revolution operation based on parthenogenesis algorithm, to enforde the local search ability. Based on MATLAB Stimulation indicated, improvement design raised the accuracy and the speed when using Genetic Algorithm to resolve VRP.(5) Leaded three-dimensional terrain into the VRP, the shortest route and the smallest gradient were objective function of multi-objective optimazation, the multi-objetive optimazation model was established, two-dimensional scan was used to creat initial population, and MATLAB was applied to multi-objective VRP.