Research On Modeling And Solving The Sustainable Supply Chain Network Design Problem

With the enhancement of public environmental awareness and sustainable development awareness,traditional supply chain management has been unable to meet the requirements of today's market environment.Only by coordinating the triple bottom line of economic efficiency,social responsibility and environmental responsibility can an enterprise keep its advantages in the fierce market competition.Therefore,it is necessary to conduct research on sustainable supply chain management.A multi-level,multi-product 0-1 mixed integer linear programming model is constructed for the sustainable supply chain network design problem,which takes into account both the economic objectives(cost minimization)of the supply chain and the environmental objectives(carbon footprint and water footprint minimization).Due to the NP-hard characteristics of the supply chain network design problem and the additional complexity added by introducing environmental indicators into the model,the exact algorithm cannot find the optimal solution for large-scale problems in an acceptable time.To address the problem,two approximate algorithms,distributed enumeration algorithm and variable neighborhood descent algorithm,are proposed in order to achieve a balance between the solution accuracy and the computing time.Distributed enumeration algorithm optimizes the value of binary variables based on the idea of enumeration,and designs a variety of filtering mechanisms to narrow the enumeration scope to approximate the optimal solution of the problem,and the algorithm is paralleled and implemented in distributed computing platform Spark to further improve the efficiency;Based on the initial solution constructed by heuristic rules,the variable neighborhood descent algorithm performs local search optimization in its multiple neighborhood structures in a determined order,and the approximate optimal solution of the problem is obtained after a number of iterations.The experimental results show that the deviation between the approximate solution obtained by the two algorithms and the optimal solution is small,and the average relative Gap of the distributed enumeration algorithm is 0.13%,and the average relative Gap of the variable neighborhood descent algorithm is 0.31%.Compared with Cplex,the distributed enumeration algorithm and the variable neighborhood descent algorithm reduce the computing time by 77%and 98%,respectively,which verifies the effectiveness and efficiency of the two algorithms,and proves that they can effectively solve the sustainable supply chain network design problemThe paper has 14 figures,10 tables,and 55 references.