Development of a multi-objective optimization model for transport and environment in a closed-loop automotive supply chain

Due to increasing severity of the environment such as limited raw materials, increasing pollutions, global warming and greenhouse gas (GHG) pollution, manufacturers have to design their supply chains to be green. For example, greening the automotive supply chain has become a major concern and car manufacturers have been obliged to recover at least 95% End-Of-Life vehicle (ELV) by 2015. Since closed-loop network make an infrastructure to collect and recover used products, developing an effective closedloop network as a major greening tool in supply chains has been growing increasingly by researchers. Most researchers have tried to minimize the total cost in logistics network and neglects environmental while the most important question in green supply chain is how to identify preferred solutions balancing environmental and business concerns. Since the logistics networks are known as complex models, exact methods could not find the optimum solution. Therefore, various meta-heuristic algorithms have been tried by researchers. In this research, a new Multi-Objective Logistics Network Model for Automotive Closed-Loop Supply Chain (MOACLSC) for recovering ELVs was developed. After that a Meta-heuristic method is used for finding the global optimum solution. The methodology of this research includes three stages. The variable (such as truck type for transportation and the amount of material that should be transshipped between facilities) and parameters (such as transportation cost, facilities’ capacity) are identified and then the conceptual model for MOACLSC was developed. In the second stage, the MOACLSC mathematical model was developed for recovering ELVs. In the last stage, an extended Gravitational Search Algorithm (GSA) as a parallel search algorithm and high convergence rate into high quality final solutions is used to solve the proposed mathematical model and to achieve the Pareto set of solution. The Multiobjective GSA (MOGSA) algorithm is adopted and then programmed using MATLAB software particularly to the MOACLSC. To verify the model, four examples from literature were considered and compared the MOGSA’s optimum solutions result by Genetic Algorithm’s (GA) result. The results obtained from problem were analyzed based on the objective function (cost), and the design parameters of the network. Analysis of results expressed the acceptable performance of MOACLSC and MOGSA compared to the proposed mathematical model in the example and GA. Comparing the total cost of the networks, revealed that the total closed-loop logistics network’s cost for ELV recovering were reduced by 1.7%, 2.4%, 3.3% and 3.9% in four problems respectively. Finally to present the model validity of a real case study in automotive industrial was studied. The result shows if the proposed model implement to redesign forward logistics, 12.36% of the total cost can be decreased. Indeed, a Pareto set of solutions including, 15 solutions were found which they can be selected a preferred solution balancing environmental and business concerns. To know the proposed model’s sensitivity, after model validation, sensitivity analysis has been done and the result has been interpreted to provide some interesting managerial insights.

Preview PDF FK 2014 25R.pdf Download (1MB) | Preview

Actions (login required)

View Item