A three-phase heuristic approach for reverse logistics network design incorporating carbon footprint

Reverse logistics (RL) is emerging as a significant area of activity for business and industry, motivated by both commercial profitability and wider environmental sustainability factors. However, planning and implementing an appropriate RL network within existing supply chains for product recovery that increases customer satisfaction, decreases overall costs, and provides a competitive advantage over other companies is complex. In the current study, we developed a mixed integer linear programming (MILP) model for a reverse logistics network design (RLND) in a multi-period setting. The RL network consists of collection centres, capacitated inspection and remanufacturing centres and customer zones to serve. Moreover, the model incorporates significant characteristics such as vehicle type selection and carbon emissions (through transportation and operations). Since the network design problems are NP-hard, we first propose a solution approach based on Benders decomposition (BD). Then, based on the structure of the problem we propose a three-phase heuristic approach. Finally, to establish the performance and robustness of the proposed solution approach, the results are compared with benchmark results obtained using CPLEX in terms of both solution quality and computational time. From the computational results, we validated that the three-phase heuristic approach performs superior to the BD and Branch &Cut approach.