Research On Facility Location And Capacity Planning Problem Considering Uncertain Demand

In recent years,the logistics industry has been developing rapidly with the rise of ecommerce and international trade.For this reason,a large number of infrastructures need to be built and the problem of such facility construction can be summarized as facility location and capacity planning problem(FLCPP).However,there are many uncertain situations in the actual location problem,which make decision-making more challenging in practice.From the current research,stochastic simulation optimization is an effective method to solve stochastic uncertain problems.Optimal computational allocation method(OCBA)has the highest accuracy and fast computing speed,but the simulation optimization method needs further research and exploration in practical application and advantages.Therefore,this thesis establishes a FLCPP model considering uncertain demand,and proposes an method called C-OCBA(CPLEX with Optimal Computing Budget Allocation)which combines the idea of traditional mathematical programming with stochastic simulation optimization.The specific research content includes the following aspects:1)Problem modeling.This thesis takes cross-border e-commerce as an example.Ecommerce platforms need to establish overseas distribution centers and consider the influence of dynamic of customer demand and the difference of national policy preferences.Then this thesis introduces the idea of stochastic programming and preference weight coefficients and establish a FLCPP model with the lowest total logistics cost.Then the idea of Monte Carlo is introduced,and the stochastic model is transformed to a linear programming model which can be directly solved by CPLEX.2)Algorithm design.The C-OCBA algorithm is proposed which combines the ideas of CPLEX and OCBA.The simulation estimates the true value by the mean of a large number of simulation results.The transformed FLCPP model is used as the simulation model.Each stage has different objective function values according to different samples.According to the mean and standard deviation of the objective function values,the simulation budget allocation for next stage is determined.A resource allocation scheme that allocates a large portion of the simulation resources to critical alternatives to improve simulation efficiency and accuracy.In other word,a larger portion of the budget should be allocated to more critical alternatives to improve simulation efficiency and accuracy.3)Simulation experiments.Comparing with two stochastic simulation optimization methods which are EA and PTV method,the experimental results show that the proposed method can better solve the problems in the model,and has higher simulation accuracy and simulation efficiency,and validates the effectiveness of C-OCBA method.Then the influence of C-OCBA parameters on the simulation results is analyzed.The results show that the reasonable selection of parameters can effectively improve the performance of the method.The method proposed in this thesis can not only be used to solve above problems,but also provide a solution to other FLCPPs,such as large-scale candidate schemes and multi-objective optimization problems,and has certain guiding significance.