| In recent years,the consumption market size of fresh products in China has exceeded 5 trillion yuan,but the negative population growth trend also leads to the weak increment of consumption end of fresh products,the demand is difficult to predict accurately,and the sinking permeability of supply end gradually tends to saturation.At the same time,various platforms adopt differentiated fresh warehouse allocation mode to compete,and the fresh product market in the future has entered the era of stock grabbing.Therefore,whether the fresh product demand can be accurately predicted,whether the warehouse allocation mode can be reasonably adjusted,and whether the demand reserves are appropriate and the location is close to the market has become the key for fresh product distribution enterprises to seek long-term development.The pre-warehouse mode is a new development trend of the fresh industry.Compared with the traditional distribution center,the pre-warehouse is usually distributed in areas with dense consumers and relatively hidden.Its characteristics of large distribution density in the city,close to consumers,low cost of site selection,easy preservation of fresh products,low loss,and relatively low cold chain distribution cost make the pre-warehouse mode can solve the problem of consumers "the last three kilometers",and can ensure the freshness of fresh goods and timely delivery rate,improve consumer satisfaction.In this paper,based on the investigation of the operation status of R fresh product distribution enterprise,aiming at the two major problems of unscientific fresh product demand forecasting method and backward and inefficient warehouse allocation mode,combining with the mainstream forecasting method and the theory of pre-warehouse mode,the demand forecasting method in line with the actual situation of R fresh product distribution enterprise and the pre-warehouse location scheme based on the pre-warehouse mode are proposed.It is necessary to re-forecast the fresh product demand of R fresh distribution enterprise according to the categories under the pre-warehouse mode,and further design the pre-warehouse location scheme according to the spatial distribution of the demand prediction results.The main research results are as follows:(1)A new demand forecasting method suitable for R fresh food distribution enterprises is provided.This paper collected the fresh purchase orders of R fresh distribution enterprise in the past five years as the original data of fresh product demand forecast.Then GM(1,1)model and ARIMA(2,1,0)model were used to forecast the demand of four fresh products in the next three years,namely,fruits,vegetables,meat and aquatic products,which meet the selected categories in the front warehouse.Through comparison and verification,the predicted results can truly reflect the future demand increment,which is helpful to solve the problem of product classification and site selection in the pre-warehouse.(2)Designed the mathematical model for the location of the forward warehouse,and gave the decision scheme for the location of the forward warehouse.The objective function of the model established in this paper is to obtain the minimum value by multiplying the sum of the values of the distribution distance between the product demand from all the selected storehouses and each cell responsible for distribution.The corresponding immune algorithm is designed by MATLAB to solve the problem,and an optimized location decision scheme is obtained.In this paper,the POI data of the four districts in the main urban area of Hengyang City in the study area were collected and analyzed by ARCGIS,and finally sorted into 80 demand points as alternative points for the pre-warehouse.Combined with the forecast results of fresh demand of R fresh distribution enterprise,the demand distribution was carried out to determine the final selection number of pre-warehouse and plug it into the model for solving.By analyzing the current situation of the business links of R fresh distribution enterprise,this paper finds out and puts forward an optimization plan for the relevant problems. |