The Vehicle Routing Problem In Network Distribution Of Low-temprature Milk

With the increasing of consumer' preferences for fresh and healthy products,domestic consumer demands for low-temperature milks are also increasing.Owing to the characteristics of short shelf life,perishable and vulnerable,the low-temperature milk need the support from cold chain logistics all along the line that both from production to sales.cold chain logistics,especially cold chain logistics distribution plays an immense role in the circulation and sales of low-temperature milk.In recent years,China cold chain logistics has grown rapidly,in the city,more and more customers need to be serviced only by a single logistics distribution center,from this the logistics distribution network becomes more complicated and confused,which bring challenges for the logistics distribution of low-temperature milk.In addition,low-temperature milk requires high timeliness,distribution radius is limited,how to arrange the vehicle route reasonably,distribute low-temperature milk fast and safely to shorten the transit time of the product,improve distribution efficiency is of great practical significance.Therefore,in this paper,we take the low-temperature milk city distribution network for study,use the two-stage algorithm of cluster-first route-second to study the vehicle routing problem of low-temperature milk in the network distribution.The main works of this thesis are as follows:(1)Research on network partitionDefined the shortest trunk distance,the actual distance instead of the European distance as the basis of network partition.At the same time,the initial center of k-medoids clustering algorithm is optimized by considering the distance and density,and the distribution quantity constraint is added.(2)VRP model construction and algorithm design for low temperature milk network distributionAccording to the characteristics of low temperature milk network distribution,the shortest path model is constructed.A new discrete cuckoo search algorithm(SA-DCS)based on simulated annealing is proposed to solve the shortcomings of the current cuckoo search algorithm.The algorithm is compared with the improved cuckoo search algorithm(RKCS)to test the validity of the proposed algorithm.(3)Application case studyY Food Co.,Ltd.is selected to distribute the real data and carry out the case study of the vehicle routing problem in low-temperature milk network distribution to further validate the validity of the model and algorithm.The results show that the improved k-medoids clustering algorithm proposed in this paper has good efficiency and robustness,and the clustering quality considering the constraint of distribution is obviously better than the clustering result without considering the constraint of distribution.And the discrete cuckoo search algorithm based on simulated annealing in this paper has a good performance in searching precision and searching performance.In addition,the two-stage method of the cluster-first route-second by the above two methods has proved to be an effective method for solving the vehicle routing problem.