Study On Vehicle Routing Problem In Physical Logistic

With the development of globalization, digitalization and integration in logistic industry, distribution plays a more and more important role in the whole physical logistic system. As the most important subsystem in distribution, transportation accounts for fifty percent of the whole expense of the logistic system. So, it must begin with the reduction of transportation cost to reduce the total cost. And whether the vehicles' routes are selected rationally or not affects directly the speed, cost and benefits of distribution. Especially, the selection of multi-customers' routes is a complexly systematic project, a rational routing plan can help achieve quicker response speed, higher quality of service, higher customers' satisfaction and lower operational cost of service merchants. Thus, VRP(Vehicle Routing Problem) has always been studied since it was presented by Danting and Ramer in 1959.VRP is normally defined as a problem that some rational routes should be organized to let one or more vehicles pass by a series of depots which deliver or receive goods in sequence, and that a certain targets such as the shortest distance, the lowest cost, the shortest time and the least vehicles should be gained under several constraints, for example, volume of demand, volume of delivery, time of delivery, capacity of vehicle, odometer of traveling, time limits, and so on. This paper does a research on VRP and contains five chapters. First, lots of relevant documents are collected, arranged and sorted. And the research status quos on VRP, especially on classic VRP, VRP with time windows(VRPTW), dynamic VRP(DVRP) and capacitated VRP(CVRP) are introduced. Then, we explain the necessity and realistic significance of solving VRP by introducing the importance of distribution in the whole physical logistic and the status quo in the home distribution industry. We put forward two mathematic models, one expressed by circles and the other by transportation cost. And then we detail different algorithms on VRP by a table. Finally, we give a genetic algorithm(GA) which is based on national number codes, chromosomes representing feasible circles, similar PMX crossover, and joining punishment in the fitness function. What's more, we prove the effectiveness of this algorithm by means of making a MATLAB program on a VRPTW example.This paper analyzes VRP mathematically and realizes it by MATLAB. The algorithm given in this paper can obtain an optimized solution effectively and has been proven to be a good one. So, it has a high value to logistic enterprises.