A memetic algorithm with a novel neighborhood search and modified solution representation for closed-loop supply chain network design

Stringent pressures from environmental requirements have forced plants to face a commencing problem, which is to improve strategies to configure forward and reverse supply chains simultaneously. With this aim, many mathematical models and solution approaches have been developed in the literature. In this paper, to tackle such problems, a memetic algorithm (MA) with an extended priority-based (EPb) encoding/decoding method based on a flexible combinatorial neighborhood search (NS) strategy is developed. Moreover, to avoid time-consuming repair process in discrete solution representation, a technique to convert the discrete representation to a continuous one is proposed. Finally, to speed up the proposed algorithm, a multi-start simulation annealing (MSA) is embedded to the MA. To assess the quality of the novel hybrid memetic algorithm (HMA), test problems from the small size for accuracy to the real size for efficiency are presented. The results are first compared with commercial solvers and then compared with other genetic algorithms (GAs) and MAs in the literature with different solution encoding/decoding methods and operators. All algorithms are applied for a closed-loop supply chain network design (CLSCND) which deals with locating facilities as well as assigning product flows and inventory costs in a multi period environment. The results demonstrate the high quality of the proposed HMA.