|
Graduation Thesis from 2018
ISSA
M.S. |
Due-date based multi-period part selection for flexible manufacturing systems with controllable processing times This study considers a multi-period part selection problem for flexible manufacturing systems in which part processing times are not given, but can be changed. The problem is to determine the set of parts and their processing times while satisfying the processing time and the tool magazine capacities in each period of a planning horizon for the objective of minimizing the sum of part processing, earliness/tardiness, subcontracting and tool costs. Some practical considerations such as the number of available tool copies, tool life restrictions and tool sharing are also considered. An integer programming model is developed to represent the problem mathematically. Then, due to the complexity of the problem, we propose two-phase heuristics in which an initial solution is obtained by a greedy algorithms under initial processing times and then it is improved using various local search methods while adjusting part processing times. To test the performances of the heuristics proposed in this study, computational experiments were done on a number of test instances, and the results are reported. |
Lee, Dong Kyu
M.S. |
Palletizing and scheduling for flexible manufacturing systems with multi-fixturing pallets This study considers a palletizing and scheduling problem for flexible manufacturing systems with multi-fixturing pallets that can load multiple parts at the same time for two distinctive objectives of minimizing the makespan and the total tardiness. The palletizing problem is to determine the set of parts to be loaded on each pallet and the resulting scheduling problem is to determine the allocation/sequence of pallets to process the loaded parts on machines. Practical constraints such as setup worker availability, re-palletizing, central buffer capacity, limited numbers of pallets and fixture type dedicated to each part type are also considered. Due to the complexity of the problem, a solution approach is proposed that decomposes the problem into the sequential sub-problems of determining the set of parts to be loaded on each pallet, the machine for each machine and solves them in sequence. To test performance of the decomposition based solution approach, simulation experiments were done on various test instances, and the results are reported. |
Kim, Hyung Won
Ph.D. |
Deterministic and Stochastic Models for Selective Disassembly Sequencing with Sequence-dependent Setups in Parallel Disassembly Environment This dissertation focuses on the selective disassembly sequencing problem with sequence-dependent setups in the parallel disassembly that removes one or more components at the same time by a single disassembly operation. The problem is to determine the sequence of disassembly operations to extract multiple target components while satisfying the precedence relations among disassembly operations. In this dissertation, we consider one deterministic model and two stochastic models for the selective disassembly sequencing problem. In Chapter 2, we suggest the deterministic model for the selective disassembly sequencing problem. The problem is to determine the sequence of disassembly operations to extract multiple target components while satisfying the precedence relations among disassembly operations. The objective is to minimize the sum of sequence-dependent setup and operation costs. An integer programming model is developed after representing all possible disassembly sequences using an extended process graph, and then an optimal branch and bound algorithm is proposed that incorporates the methods to obtain the lower and upper bounds as well as a dominance property to reduce the search space. To show the performance of the algorithm, computational experiments are done on various random vi instances, and the results are reported. In particular, it is shown from the test results that the proposed algorithm requires much shorter computation times than a competitive commercial software package. Finally, a case is reported to illustrate the extended process graph and the solution algorithm. In Chapter 3, we propose two stochastic models for the selective disassembly sequencing problem. The first one is the problem that considers random operation times in the parallel disassembly environment in which one or more components can be removed at the same time by a single disassembly operation. An extension of the first one, the second one is the problem that considers irregular disassembly operations and random operation times. Here, the irregular operations are defined as those in which fasteners can be removed by additional destructive operations without damaging to the target components. After representing all possible sequences using the extended process graph, two stochastic integer programming models are developed that minimizes the sum of disassembly and penalty costs, where the disassembly cost consists of sequence-dependent setup and operation costs and the penalty cost is the expectation of the costs incurred when the total disassembly time exceeds a given threshold value. A sample average approximation algorithm is proposed that incorporates an optimal algorithm to solve the sample average approximating problem under a given set of scenarios for each problem. The algorithm is illustrated with a hand-light case and a large sized random instance, and the results are reported. |
Yu, Jae Min
Ph.D. |
Due-date based Scheduling Algorithms for Job-shop-type Remanufacturing Systems with Component Matching Requirement This dissertation focuses on a scheduling problem for remanufacturing systems with parallel disassembly workstations, a job-shop-type reprocessing shop and parallel reassembly workstations, in which the components obtained by disassembling an end-of-use/life product must be matched when reassembling the corresponding remanufactured product, i.e. component matching requirement. In this dissertation, we first consider a scheduling problem for job-shop-type reprocessing shop with component matching requirement. Then, a scheduling problem for job-shop-type reprocessing shop and parallel reassembling workstations is considered. Finally, a scheduling problem for the entire remanufacturing system composed of parallel disassembly 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 requirement iv is considered. To consider the component matching requirement, we propose 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 duedate based objective of minimizing the total family tardiness, 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 practical 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 experiments 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 reassembly 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. |
|
