Research On The Optimization Of Reverse Logistics Network Of End-of-life Vehicles In Uncertain Environment

In recent years,due to the influence of many factors such as economic benefits,government legislation,environmental impacts,and social responsibility etc.,the government and manufacturers pay more and more attention to the reverse logistics of end-of-life vehicles.To establish the reverse logistics network system of end-of-life vehicles and promote the remanufactured automobile products have become an inevitable way for the sustainable development of China’s automobile industry.Based on the economic,environmental and social perspectives,it is of great theoretical and practical significance to study the optimization of the reverse logistics network of end-of-life vehicles in the uncertain environment.The purpose is to help automobile manufacturers to plan the reverse logistics network scientifically and rationally under the uncertain environment,so as to achieve the maximum economic and social benefits with meeting the requirements of environmental legislation.Firstly,based on the current research status,a five-layer reverse logistics network with the functions of processing,remanufacturing,reusing and redistributing is constructed,and a multi-objective model mixed integer linear programming is established to optimize the reverse logistics network of end-of-life vehicles in a certain environment by considering the economic cost,environmental impact and social benefits.Secondly,taking uncertain factors such as recovery rate,transportation cost,processing cost and remanufacturing cost into account,uncertain set and robust optimization model of the reverse logistics network optimization of end-of-life vehicles in the uncertain environment are built.Thirdly,the optimization problem of the reverse logistics network for end-of-life vehicles is divided into the discrete facility location problem(FLP)with capacity limitation and the vehicle routing problem(VRP).Using genetic operator improvement,specific problem design and algorithm parameters set improvement methods,the genetic algorithm(GA)combining elite strategy with roulette and the GA based on priority coding are designed to solve the FLP problem and VRP problem respectively.Finally,the cases under deterministic and uncertain environment,and the influence cases of uncertain factors are designed and solved by CPLEX software and GA.The results show that the accuracy and efficiency of CPLEX is better than GA in solving small-scale cases,and it is suitable for the network with less than 80 customers.The advantage of GA for large-scale cases is obvious,and suitable for networks with more than 80 customers.The robust optimization model in the uncertain environment considers both the minimum logistics cost and the robustness of the network,which could better resist theinternal and external risks of the network system.Processing and remanufacturing cost have the greatest impact on network cost,followed by recovery rate,and the transportation cost has the least impact.It shows that the built optimization model is reasonable and the designed algorithm is scientific and effective,this provides a theoretical basis for automobile manufacturers to reduce the risk of uncertainty and build reverse logistics network decision-making.