Research On The Model Of The Location Of Emergency Supplies Repository Under Demand Uncertainty

In recent years,all kinds of emergencies occur frequently,which caused a serious threat to the people's life.In order to improve the prevention and the ability of emergency incident,strengthen the construction of emergency supplies reserve,and ensure emergency demand of the goods.On the basis of predecessors,there are three questions to solve about the uncertainty demand of the condition of emergency supplies,main: demand forecasts,the selection of the objective function,the determination of solving method.To settle these questions,the article builds a multi-objective decision-making model of inventory.Then,the article makes an empirical analysis about Hefei.This paper focuses on where to build the inventory before the emergency supplies happened,which belongs to the early prevention stage.The location of question integrated the management science,statistics and modeling simulation theory knowledge,using qualitative and quantitative analyses,from multiple perspectives to consider the location problem.Firstly,before solving the problem of the inventory,we need make a forecast about the demand of emergency supplies.Normally,there are many factors to influence the demand,so this paper considers the demand under the uncertainty.Through reading the literature and relevant information,there are three ways to deal with demand uncertainty mainly: first is to the robust optimization of the model uncertainty;second is to use fuzzy number uncertainty;third is to use contact number processing uncertainty.Among them,the third way has an advantage to settle with the less sample data,achieving more accurate prediction effect,which is applicable to the shortage of emergency data,so this paper adopts the method of to forecast the demand.Prediction method is to turn the demand into interval number,then add grey factors as interval number,getting the prediction results based on connection number interval grey number.Secondly,the traditional location models are summarized,contrast and researched,building a multi-objective model of emergency material,which contains emergency response time,emergency supplies to cover degree and the reserve cost,considered three aspects of site selection model.Then comparing the solving method of multi-objective function,the model of this paper is using the improved NSGA II to solve.Compared with the traditional NSGA II algorithm,the improved method combined with the finite difference method,improving the understanding of distribution,insuring the accuracy and understanding of diversity.Finally,make an empirical analysis to the Hefei as the research object.According to the population,economic level and traffic conditions in different regions in Hefei,this paper insight needs weight.Then using grey number predicts method to forecast the demand,and carries on the model.The results show that :a gray number based on connection number prediction model has higher accuracy.The three objectives are mutual connection and constraints,and DE-NSGA II method is better than that of NSGA II on time,improving the distribution of results.