Research On Optimization Model And Algorithm For Multi-Echelon Location-Routing Problem Of Logistics Network

The design of the logistics network influences directly the logistics cost of enterprises.As infrastructure of logistics system,the flaw of it cannot be amended in the later management and operation.With the popularization of E-commerce and the development of economic globalization,the scope of logistics activity continues to expand and the logistics system has become more hierarchical.The study on location-routing problem(LRP),the integration of facility location problem(FLP)and vehicle routing problem(VRP),is the reseach trend of the optimization of logostic network,and is also a hot topic of the logistic research fields.Therefore,the thesis studies the optimization of multi-echelon location-routing problem(ME-LRP)in logistics distribution network by using System Engineering Theory and its methods under the guidance of Graph Thoery and Complex Network Theory.The main research contains following parts:First,a mathematical model is built to study ME-LRP.Based on the knowledge of logistics distribution networking,the thesis takes 3E-LRP as the study cut-in point,which is abstracted into a complex network displayed by digraph.Considering various technical andeconomic indicators such as facility capacity,vehicle type,car load and distance,the thesis builds a mathematical model of 3E-LRP that comprises two echelon of location of facility and three echelon of vehicle route planning.Under such circumstances,the hierarchical structure of logistics network is abstracted into N(N ?3)echelon,and the general mathematical model that can contain all intermediate echelons of location of facilities and all echelons of vehicle route planning is built.The complexity of model is aslo analyzed.Second,a systematic optimization between FLP and VRP is studied.According to the relationships among location of facilties,route planning and location-route,the problem of location-route is devided into three sub-decision problems,namely the location of facilities,facility allocation problem(FAP)and route planning respectively.Based on their inner relationships,the thesis proposes a new solution that integrates the Quantum Evolutionary Algorithm(QEA)and Genetic Algorithm(GA)to guarantee the rationality of the global optimization of location-route problem.The thesis verifies the optimization plan and the feasibility and validity of the algorithm by testing and analyzing the simulation example of 3E-LRP.Third,a general structured ME-LRPe algorithm applicable to any layer is designed.Based on ME-LRP general model and the optimized integration of FLP with VRP,the thesis proposes a general structured algorithm based on QEA-GA master-slave mode of ME-LRP.The thesis also analyzes the space complexity from the perspectives of ME-LRP mathematical model and solution.Fourth,in order to improve the optimation efficiency of ME-LRP,a FAP optimization strategy is proposed based on the search strategy of distribution region and taking into account the weight of the shortest path.The thesis also designs and analyzes the optimal search strategy.Besides,the thesis proves the effectiveness in narrowing search space,improving search efficiency,and enhancing quality through simulation test.Fifth,a new computing method for VRP cost based on the cost coefficient of no-load vehicles is proposed.Considering the basic hypothesis of LRP,and the cost of back track of no-load vehicles and the load factor during the distribution,the thesis designes a new VRP expanses assessment formula,and deduces and proves the computational formula related to cost coefficient of no-load vehicles.Based on Graph Theory and Complex Network Theory,a general mathematical model for location-route planning of multi-echelon logistics network is built.A general structured solution for ME-LRP is proposed,which realizes the collaborative optimization of facility location and path planning at system level.The simulation test results show that the model and algorithm presented in this paper are feasible and effective.This paper establishes a general mathematical model of ME-LRP for the first time,with the lowest total cost of the system as the goal,and taking into consideration a variety of factors such as facility capacity and transport vehicles.Based on the systematic integration of FLP and VRP,the model covers all intermediate level facilities location selections and all levels of vehicle route planning,and theoretically provides a research basis for ME-LRP.The QEA-GA based general ME-LRP algorithm provides a theoretical ground f for ME-LRP research and an algorithm for the design,planning,and application of logistics network.The research in this paper fills in the deficiency of current ME-LRP research,and its practical significance is to provide theoretical and methodological guidance for the planning and design of the top level of logistics network and the expansion,upgrading and optimization of the existing logistics system.