Research On The Multi-objective Vehicle Routing Optimization In Urban Express Distribution

In recent years, the domestic e xpress industr y develops rapid ly, meanwhile, it a lso suffers fro m t he pressure of higher service qualit y custo mers require, espec ially w it h t he continuo us increase of cons umer demand and effic iency awareness. Urban express distr ibut ion, as the fina l part of the cour ier service, is the last mile in t he process of express transport ation, which p lays a key role in the improve ment of corporate service leve l. Therefore, how to achieve effic ient delivery has become an important issue for the courier companies. As the fina l result of express distr ibut ion is us ually determined by the arrange ment of ve hic le route, it is necessary and meaningful to study the routing optimizat ion of the delivery vehicles.According to the status of t he domest ic and internat iona l research, this paper focused on the proble m of mult i-objective vehic le routing op timizat ion in urban express distr ibut io n. First ly, based on the characterist ics of urban express distrib ut ion, we ma inly took the factors of vehic le type and uncertain t ime into cons ideration, and then constructed a model of mult i-object vehic le rout ing problem w it h t ime w indows, whic h aimed at achieving t he goals of the minimum total cost, the lowest time punis hment and the least vehic les required. Next, we adopted a kind of improved genetic a lgorit hm to solve t he model. A mong t he improved genet ic algor it h m, w ith regard to the design of fit ness function, we used the met hod of unifor m dimens ion and weighted ratio in order to trans form the mult i-objective funct ion into a single- objective funct ion. Moreover, w ith respect to the design of genet ic operators, we used the mechanis m of saving t he best individua l a mong the populat ion in the process of select ion, crossover and mutatio n genet ic operatio n, whic h could improve the overall leve l of populat ion evolut ion and get the approximate optima l solut ion. In the last, we combined wit h document examp les to verify this a lgor ithm. In addit ion, we also ana lyzed the influence of the re levant factors on the results of t he experiment by adopting numerical examples.According to the exper imenta l results, the optima l results applying the improved genet ic algor it hm were better than the experime ntal group so that the valid it y of the genet ic algorit hm could be verified well. In addit ion, we also tested the impact of the rele vant factors on the experimenta l results. The results showed that the time w indow factor had an important influe nce on the result of route optimizat ion, and the mult iple types had advanta ges in the path optimizat ion compared with the single vehic le type. Therefore, it could improve the service leve l by reducing t ime w indow pena lt ies and cut down t he distribution costs by increasing the types of vehicles for the express enterprise s.