부품 매칭 제약을 갖는 개별공정 형태의 재제조 시스템에 대한 납기 기반 스케줄링 알고리즘

Abstract

Due-date based Scheduling Algorithms for Job-shop-type Remanufacturing Systems with Component Matching Requirement Yu, Jae-Min Department of Industrial Engineering Graduated School of Hanyang Univer-sity This dissertation focuses on a scheduling problem for remanufacturing systems with parallel disassembly workstations, a job-shop-type reprocessing shop and paral-lel reassembly workstations, in which the components obtained by disassembling an end-of-use/life product must be matched when reassembling the corresponding re-manufactured product, i.e. component matching requirement. In this dissertation, we first consider a scheduling problem for job-shop-type reprocessing shop with com-ponent matching requirement. Then, a scheduling problem for job-shop-type repro-cessing shop and parallel reassembling workstations is considered. Finally, a sched-uling problem for the entire remanufacturing system composed of parallel disas-sembly workstations, a job-shop-type reprocessing shop and parallel reassembly workstations is considered. In Chapter 2, a job-shop-type reprocessing shop with component matching re-quirement is considered. To consider the component matching requirement, we pro-pose a job shop scheduling problem in which jobs are grouped into job families, but they are processed individually using their distinct routings. Unlike the previous studies, we consider a due-date based objective of minimizing the total family tardi-ness, i.e. sum of positive deviations between the due-dates and the completion times of job families. A mixed integer programming model is developed to present the problem mathematically. Then, an optimal algorithm is proposed using the branch and bound technique while developing a job family based lower bound. For practi-cal applications even with the large-sized job instances, two types of heuristics, shifting bottleneck based and priority scheduling algorithms, are also proposed. To test the performances of the three types of solution algorithms, computational exper-iments were done on various test instances and the results were reported. In Chapter 3, we consider a scheduling problem of determining the sequence of the jobs to be processed on each workstation of job-shop-type reprocessing shop and the allocation/sequence of the jobs to be performed on parallel reassembly workstations. To cope with component matching requirement, the reprocessing jobs are grouped into job families corresponding to products to be remanufactured. A mixed integer programming model is proposed for the objective of minimizing the total tardiness. Then, two solution approaches, sequential and integrated ones, are proposed. The intuitive sequential approach solves the reprocessing and the reas-sembly scheduling sub-problems in sequence while the integrated approach solves them simultaneously after representing the problem as an extended disjunctive graph. It is shown from computational tests that the integrated one outperforms the intuitive sequential one significantly. Also, the absolute performance of the better integrated approach is reported using the gaps from optimal solution values for small sized test instances. In Chapter 4, a scheduling problem of an entire remanufacturing system is con-sidered. In this chapter, an entire remanufacturing system is defined with parallel disassembly workstations, a job-shop-type reprocessing shop and parallel reassem-bly workstations, in which the components obtained by disassembling an end-of-use/life product must be matched when reassembling the corresponding remanufac-tured product, i.e. component matching requirement. The problem is to determine the allocation/sequence of jobs on the parallel disassembly workstations, the se-quence of the jobs on each workstation of job-shop-type reprocessing shop and the allocation/sequence of on the parallel reassembly workstations. To present the com-ponent matching requirement, the reprocessing jobs are grouped into job families each of which corresponds to a product to be remanufactured. After an integer pro-gramming model is developed, two types of solution algorithms, decomposed and integrated ones, are proposed, where the decomposed ones solve the disassembly, reprocessing and reassembly scheduling sub-problems separately while the integrat-ed ones solve them at the same time after representing the problem as an extended disjunctive graph. Computational experiments were done on a number of test in-stances and the results show that the integrated algorithms outperform the intuitive decomposed ones significantly.