Model And Algorithm Of Vehicle Routing Problem With Stochastic Delivery Time

The vehicle routing problem is a key problem of the logistics distribution optimizationwhich has always been the hot topic in the modern logistics field. Effective vehicle routingcan not only improve the response speed of the customer demand and the service quality butalso reduce operational cost of logistics service provider. Now the research for VRPconstaints such as traveling time and service time regards as a fixed static problems, and thetarget multi-function model set starting from the distribution center. The model objectivefunction is set such as the shortest distance and the lowest cost, the study about consideringthe customer and distribution costs multi-objective optimization is rare. But in the actualdistribution of traffic due to vehicles and natural conditions impact, so that delivery vehicleshave a certain randomness in the distribution process for the customer. The study of VRP witha random distribution is more close to the practical situation.The main idea of thesis is studying how to optimize the route of ration. By analyzing theexisting study, it points out the problems to reduce the main idea of thesis. Based on studieson how to optimize the route of ration. The thesis establishes a new feasible mode ofoptimizing the route of ration, and validates the feasibility of the mode by an example.Firstly, we introduce the concept of VRP. This dissertation is based on various factors ofenterprise distribution cost minimization and customer satisfaction maximization, andresearches the fuzzy characteristic of logistics distribution arrival time and delivering time ofthe randomness and customer requirements. By random chance constrained programmingtheory establish stochastic chance constrained programming model, the mode considerscustomer’s satisfaction as a restriction, which has the direct effect. It quantifies the lost ofcredit indirectly, so as to improve the long benenfit of the ration center and upgrade thecustomer service, punctuality and high efficicence.Secondly, the application of genetic algorithms in the process of solving the model. Thestandard genetic algorithm is applied into vehicle routing problem with common defects ofearly convergence and easily falling into local minima. In this thesis, based on the model ofVRP characteristics. Traditional genetic algorithm is improved and a new adaptive geneticalgorithm is introduced. That is, the new algorithm can adaptively adjust the probabilities ofcrossover and mutation according to the fitness values in the different evolution stages.Finally, the effectiveness and feasibility of the improved method is verified bycalculating the examples. Establish a random VRP model for examples. Adaptive geneticalgorithm is used to solve the model. The impact on the modle solution by differentconfidence and satisfaction value is discussed. This research not only has directive significance for the practical applications of vehicle routing problem but also providesdecision support for logistics distribution scheduling system.