Researching On The Layout Of Express Service Network Based On Clustering Algorithm

The supply of goods or logistic services should be more people-oriented, serviceoriented, and the demands for services are diversified in the modern logistics, the distribution center of logistics as a symbol of modern logistics. Rational composed solution can effectively reduce unnecessary transportation costs and reduce the flow of goods in region, what's more, it also can help enterprises to improve production efficiency, reduce inventory, increase the change amount per unit time inventory goods.The center of logistics will go through next several steps: access shipping, transporting, storage, distribution and other operational processes, but the implementation of these processes are built on the basis of the network, so if you want to have no increase on the premise of the cost by selecting reasonable service shop layout programs to have a positive impact on the overall efficiency of distribution operations courier company. In this case, the dissertation location model is as a starting point, combined with the user distribution of the underlying trend, presented a potential purchase intent based on user clustering algorithm k-means location model, and its optimum solved courier network site layout policy.From the perspective of theoretical analysis point of view, in a multi-target model based on the location of the point of demand, the k-means clustering algorithm, ant colony algorithm applied to the network layout of the site, in order to better use of this particular case models for k-means clustering algorithm and ant colony the following improvements:(1) between the k-means clustering algorithm in each iteration, the definition of a balanced value, this value is a measure of equilibrium each work area assignments are balanced indicators, to ensure the final clustering A cluster is evenly distributed in each center nearby coordinates;(2)improved ant colony algorithm for path selection mainly by reduce the amount of calculation, reduce the number of calculation and rout using a balanced binary tree ant colony optimization routing three ways to optimize the process of large-scale data, the lower efficiency of classical algorithm, and avoid to fall into local optimal solution in the pathfinding process. Wherein the improvement of the k-means algorithm although the efficiency and classical k-means algorithm, a slight decrease, but it does avoid the classical k-means algorithm may unevenly distributed in various clades problem internal set point, and on the UCI data sets algorithm simulation and analysis, in order to verify the performance and effectiveness of clustering algorithms.Besides that, through in-depth analysis of vehicle scheduling model, summed up the impact of vehicle scheduling efficiency influencing factors. At the same time the ant colony algorithm, the simulation algorithm TSPLIB database as a data source, several aspects of the optimal number of hits, average and best time evolution curve simulation improved ant colony algorithm from optimal algorithm, the average solution, and classical colony algorithm, by comparing them to prove that the improved algorithm in the above aspects have improved significantly, especially for large-scale data entry, improved ant colony algorithm has time complexity; while improving ant colony algorithm can avoid the local optimal solution. Finally, through the introduction of constraint weights, expansion bottleneck, dynamic demand satisfaction and adaptability program indicators compared with the cross median model, maximum coverage model and the advantages and disadvantages of discrete point coverage heuristic location model three kinds of location model, location draw discrete point coverage heuristic model to meet the dynamic needs of users, scalability, and so on has a better performance.Combined with these improved k-means clustering algorithm and ant colony, apply them to the courier shop layout model, the simulation needs of the regional network in Chengdu area. Analog outlets coordinate within the region as the input data sample clustering analysis of "social" feature that the user purchase, provide a theoretical basis for the next location. Then the user's cluster analysis of the "social" characteristics of the user, and then carved out a preliminary layout area. After discrete point coverage heuristic model to determine the final location of the site plan, and give the scope of radiation each site service center and best distribution route in each service area.Based on the delivery model analysis, optimization, simulation and analysis of simulation results show that reasonable network layout strategy without increasing the cost of inputs, the logistics and distribution process, to improve the operational efficiency of logistics and distribution systems are improved significantly, which played a positive role and did good impacted to the development of modern logistics industry and logistics service quality.