Statistical Inference For Location Problem Of National Large Gatherings

This paper raised a brand new location problem.We want to determine the best location of national large gatherings.This new problem is different from the past location problems of conferences,for the numbers of conference participants are certain and available,so only need to build the lowest cost model.On the contrary,the numbers of activity participants are uncertain because people usually make the decision by the location of activity.This problem is also different from the sptial competition with price problem,for the amount of alternative stores is large,but the large gathings in this problem is the only one in the period and it could not be replaced.In order to hold the best large gatherings,in other words,to get the most participants,we build a new location model to solve this problem.In order to estimate the amount of participants,we brought in the concept of partici-pating determiner,which is related to the cost of participating the activity and the national economic condition.In order to maximize the amount of participants,we collect the train fare between all national provinces,and we calculate the least cost with the help of Floyd's algorithm.In order to get the more accurate participating determiner,we collected nation-al economic condition and did cluster analysis.The participating determiners of different provinces change with the cost.We first did questionnaire survey so as to solve the model.We got the estimator by analyzing the results of survey.After programming in MATLAB,we choose the preferred destinations of the basic model and the transit fee added model.In chapter 4,we considered the constraints of time and the other cost besides round trip cost,including local cost and the admission tickets.We also designed a tour routine.Finally,we programmed in R to make a population forecast so as to determine the future destination.