Research On Model And Algorithm For The Multi-commodity And Multi-source Closed-Loop Location-Inventory-Routing Problem

With the rapid development of information technology,e-commerce and global economic integration,customer demand presents the characteristics of diversity,uncertainty and dynamics,and the phenomenon of goods return is also very common.In order to meet the needs of the development of the times,logistics enterprises need to constantly improve the level of logistics services,improve the efficiency of logistics system operation,reduce the operation cost of logistics system,and realize the rational allocation of resources.The location of facilities at strategic level,inventory control at tactical level and route planning of distribution and recycling are three key problems in the operation and practice of logistics system.In addition,in the Internet era,commodity trading has shown the characteristics of small demand,abundant types,scattered locations,increasingly prominent phenomenon of commodity return,and high requirements for logistics services.The forward distribution and reverse recovery of goods with various demands put forward new requirements for integrated planning of logistics system.According to this practical background,this paper studies the new location-routing problem,location-inventory problem and location-inventory-routingproblem respectively.The main research results are summarized as follows:(1)The mathematic model for multi-commodity and multi-source closed loop locating-routing problemconsidering return is established.The objective is to minimize the sum of the construction cost,forward logistics distribution cost and reverse logistics recovery cost of the distribution and recycling center.In this problem,the location decision of distribution recycling center location decisionand distribution recycling routingdecision of the customers should be made.Based on the analysis of theL characteristics of the investigatedproblem,a greedy heuristic algorithm is proposed to quickly construct the initial feasible solution.Based on the eight effective neighborhood structures proposed to optimize location decision and the optimal routing decision,an efficient variable neighborhood search algorithm is designed.The experimental results show that the variable neighborhood search algorithm proposed in this papercan obtain the optimal solution or approximate optimal solution for small scale instancesin a relatively short time,and obtain the satisfactory solution for largescale instances,and it is obviously superior to the optimization software CPLEX in solutionquality and solutiontime.(2)The mathematic model for multi-commodity and multi-source closed loop locating-inventory problemconsidering return is established.The objective is to minimize the sum of the construction cost,ordering cost and inventory cost of the distribution and recycling center.In this problem,the location decision of distribution recycling center location decisiona and order quantity decision of each commodity should be made.The hybrid differential evolution algorithm is designed,and the performance of the algorithm is adaptedby improving the crossover operation,designing the mutation operator,introducing the truncation selection strategy of genetic algorithm and changing the termination condition.Experimental results show that the proposed algorithm has good optimization ability and is superior to the existing differential evolution algorithm in the related literature and the optimization software LINGO in terms of solution qualitity and solution time.(3)The mathematic model for multi-commodity and multi-source closed loop locating-inventory-routing problemconsidering return is established.The objective is to minimize the sum of the construction cost,ordering cost,inventory cost,forward logistics distribution cost and reverse logistics recovery costof the distribution and recycling center.In this problem,split delivery and split pickup for delivery and recovery requirements at each customer are allowed,the location decision of distribution recycling center location decision,order quantity decision of each commodity and distribution recyclingroutingdecision of the customers should be made.First,based on the analysis of the characteristics of the problem,the original problem is transformed into a two-stage decision problem by firstlyconsidering the location decision of the distribution recycling center and the route decision of the customers,then the order quantity decision of each commodity in the distribution recycling center,and then the two-stage heuristic algorithm is designed to find the local optimal solution.An iterative local search algorithm is designed to find the global optimal solution by simultaneously considering the location decision of the distribution and recycling center,the order quantity decision of each commodity and the distribution and recycling routingdecision of customers.Experimental results show that the iterated local search algorithm is designed based on global optimal isobviously superior to the two-phase heuristic algorithmis designed based on local optimal.In addition,by comparing with the results provided by optimization software LINGO,we vertify that the iterated local search algorithm is designed in this paper canquickly obtainthe optimal solution or approximate optimal solution.