Supplier selection and purchase problem for multi-echelon defective supply chain system with stochastic demand

In the real production process, some members in the supply chain system sometimes cannot effectively complete their production task because of defects involving the production or purchasing of components. A supply chain system that has defects in at least one echelon is called a multi-echelon defective supply chain (MDSC) system. Most supply chain systems are MDSC systems. Determining parts or components supply quota from different suppliers with limited suppliers, factories and distribution centers capacities in the supply chain system are becoming an important issue for businesses. In this study, we propose a new heuristic (H2) which is an extension of H1 heuristic that was previously presented. The MDSC system was formed with the mixed integer linear programming by LINDO software for calculation of the lower bound. The heuristics and MDSC system were modeled by using ProModel software. The heuristics were applied to a case from the Turkish furniture industry. The heuristics were compared with each other by considering different coefficients of variation, service levels, and deviation from lower bound. Simulation experiments showed that the proposed H2 heuristic outperformed the H1 heuristic.In the real production process, some membersin the supply chain system sometimes cannot effectivelycomplete their production task because of defectsinvolving the production or purchasing of components. Asupply chain system that has defects in at least oneechelon is called a multi-echelon defective supply chain(MDSC) system. Most supply chain systems are MDSCsystems. Determining parts or components supply quotafrom different suppliers with limited suppliers, factoriesand distribution centers capacities in the supply chainsystem are becoming an important issue for businesses. Inthis study, we propose a new heuristic (H2) which is anextension of H1 heuristic that was previously presented.The MDSC system was formed with the mixed integerlinear programming by LINDO software for calculation ofthe lower bound. The heuristics and MDSC system weremodeled by using ProModel software. The heuristics wereapplied to a case from the Turkish furniture industry. Theheuristics were compared with each other by consideringdifferent coefficients of variation, service levels, anddeviation from lower bound. Simulation experimentsshowed that the proposed H2 heuristic outperformed theH1 heuristic