Study On Site Selection Of Fresh Distribution Center Based On Spatio-temporal Data Mining

With the continuous improvement of our country's economy and living standards,fresh products have become one of the essential daily necessities in people's lives.It is an important source of vitamins and trace elements in human body,and its quality problems are gaining more and more attention.Fresh products with short shelf life,perishable and other characteristics,in the transport process generally use refrigerated trucks,that is,it is need to be transported in cold chain conditions.In order to ensure that customers can buy fresh products,most enterprises and fresh retail outlets establish their own fresh product distribution center to unified distribution fresh products.Therefore,the location of the fresh distribution center is directly related to the profitability and development of the enterprise or fresh retail outlets.How to choose the location of the fresh distribution center scientifically and rationally is the key problem of this paper.First of all,the article describes the basic theory of fresh distribution center,fresh life cycle,data mining,data cleaning and distribution center location.The function and type of fresh distribution center are explained in detail.This paper expounds the basis of the content of the fresh product life cycle function,introduces the principle of data cleaning,the main method of data mining and general process,describes the distribution center location principle,influence factors and steps,etc.These basic research for the data of fresh orders mining and fresh distribution center location model to create provides the theoretical support.Secondly,the steps and methods of data cleaning are discoursed.The algorithm of spatial clustering analysis and time feature analysis are discussed.On this basis,the use of non-equal coverage radius model to determine the distribution center of the candidate can be built area.And then use the planning method to build the site model with the minimum cost objective function of the fresh delivery system,which includes several costs such as energy consumption,facility cost,land cost,transportation cost and loss caused by freshness reduction,and put forward Model solving method.Finally,the application of the analysis and solution.First of all,cleaning and digging the order data of 020(Online To Offline)platform provided by Shijiazhuang fresh sales network,analyzes the change of customer demand in different time periods and characteristics of customer demand in different regions by spatial clustering analysis and time feature analysis.By constructing the location optimization model of the fresh distribution center,combining with the data obtained from the data mining and the data obtained from the field research,the final result of the example is solved by using the programming solution of the non-equal coverage model and the minimum cost objective function.Which verifies the feasibility of the location model.And then the existing fresh outlets for analysis and evaluation,so as to give fresh sales outlets to provide fresh purchase of the decision support,to avoid the distribution center inventory unreasonable situation,thereby reducing unnecessary cost loss.The results of this study provide a reference to the freshness of the network from both theoretical and practical aspects.This article is based on 020 big data quantitative analysis,with time and space characteristics and customer demand characteristics of temporal and spatial multidimensional feature location model is constructed and has certain innovation.However,There is also slightly insufficient,fresh corruption function and calculation of site selection model still need further detailed research.