Dynamic Lot-sizing Algorithms for Remanufacturing Systems with Multi-level Structured Products and Components Commonality

Abstract

This dissertation focuses on a lot-sizing problem for remanufacturing systems with a single disassembly workstation, unrelated parallel reprocessing workstations and a single reassembly workstation, in which end-of-use/life products are disassembled into subassemblies/parts on a disassembly workstation, each recoverable part is reprocessed on one of parallel reprocessing workstations and the reprocessed parts, with newly purchased ones due to defectives, are reassembled into remanufactured products on a reassembly workstation. In this dissertation, we first consider a lot-sizing problem for disassembly and reprocessing workstation with the multi-level structured product. Then, a dynamic lot-sizing problem for the entire remanufacturing system with multi-level structured products is considered. In Chapter 2, we consider a dynamic lot-sizing problem for remanufacturing systems in which end-of-use/life products are disassembled into their parts on a single disassembly workstation and then each part is reprocessed on one of parallel reprocessing workstations. The problem, which integrates disassembly and reprocessing lot-sizing for multi-level disassembly structures, determines disassembly and reprocessing lot-sizes as well as the workstation on which each part is reprocessed while satisfying workstation capacities and reprocessed part demands in each period of a planning horizon. An integer programming model is developed that minimizes the sum of setup, operation and inventory costs. After the NP-hardness is shown, two-stage heuristics are proposed in which an initial solution is obtained and then improved by full and partial bi-directional moves. Computational results show that the full move-based heuristic gives near-optimal solutions for small-sized instances and the partial move-based heuristic gives fast solutions with appropriate qualities. In Chapter 3, we consider a dynamic lot-sizing problem for remanufacturing systems in which end-of-use/life products are disassembled into subassemblies/parts on a disassembly workstation, each recoverable part is reprocessed on one of parallel reprocessing workstations and the reprocessed parts, together with newly purchased ones due to defectives, are reassembled into remanufactured products on a reassembly workstation. In particular, the previous studies are extended to the one for multi-level structured products with components commonality. The problem is to determine the disassembly, reprocessing, and reassembly lot-sizes as well as the workstation to reprocess each part in each period of a planning horizon while satisfying remanufactured product demands and workstation capacities. A mixed-integer programming model is developed for the objective of minimizing the sum of setup, operation and inventory holding costs. After the NP-hardness is shown, two-stage heuristics are proposed that consist of an initial solution and improvement by bi-directional moves, i.e., individual and cluster moves. Computational results show that the cluster bi-directional movement-based heuristic that considers common components outperforms the individual one and gives near-optimal solutions for small-sized test instances. Unlike the previous lot-sizing problems for remanufacturing systems, this dissertation considers the multi-level structured product and components commonality, which made the system into a divergent and convergent structure and a complex structure that makes decision-making difficult. To address the problem, heuristic algorithms were developed by reflecting the points to be added one by one for each chapter, and the performance of the solution was investigated through relative comparison.