Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/83511
Title: | 零工式生產系統動態派工與預防保養研究 Dynamic dispatching and preventive maintenance in job shop production |
Authors: | Bing-Heng Wu 吳秉恆 |
Advisor: | 吳政鴻(Cheng-Hung Wu) 吳政鴻(Cheng-Hung Wu | wuchn@ntu.edu.tw | ), |
Keyword: | 動態派工,預防保養,產品多工序, Dynamic Dispatching,Preventive Maintenance,Product multi-process, |
Publication Year : | 2022 |
Degree: | 碩士 |
Abstract: | 本研究提出應用於多產品多工序多機台之最佳動態派工與預防保養決策框架。在過往的多產品單一工序系統之派工與預防保養最佳化上,進一步延伸至多工序概念。多工序是當今日益複雜的生產製造系統經常需要考量與面對的問題,本研究針對此議題上深入研究。 本研究針對該多產品多工序多機台之生產系統,首先以混合整數線性規劃法(MILPD)進行產能分配,達到大問題分解之效果,進而降低最佳化決策求解時間。接著利用馬可夫決策過程(Markov decision process)與動態規劃法進行各生產機台與作業之動態派工與預防保養決策求解。 本研究針對提出的解決方法,進行數值實驗模擬驗證,並與其他常見派工方法進行性能對比,包含平均週期時間與總產出。實驗結果表明,本研究方法能提供相較於其他常見派工方法更優異的性能。藉由統計顯著性分析與產能利用率之敏感度分析,表明在系統產能利用率高的場合,能提供具明顯優勢的性能表現。而在系統產能利用率較低的場合,依然能提供優於或至少不差於其他常見派工方法之性能。 本研究續提出未來研究方向,本研究能做為日後更複雜生產系統派工與預保養最佳化的研究基礎與依據。 This research proposes an optimal dynamic dispatch and preventive maintenance decision-making framework for multi-product, multi-process, and multi-machine systems. In the past multi-product single-process system optimization of dispatch and preventive maintenance, it is further extended to the multi-process concept. Multi-process is a problem that often needs to be considered and faced in today's increasingly complex manufacturing systems. This research focuses on this topic. For this multi-product, multi-process, and multi-machine production system, this research first uses Mixed Integer Linear Programming Decomposition (MILPD) method to allocate capacity to achieve the effect of decomposing large problems, thereby reducing the optimal decision-making time. Then use Markov decision process and dynamic programming method to solve the dynamic dispatch and preventive maintenance decision of each subsystem. In this study, the proposed solution is verified by numerical experiment simulation, and the performance is compared with other common dispatch methods, including the average cycle time and throughput. The experimental results show that the proposed method can provide better performance than other common dispatching methods. Through statistical significance analysis and sensitivity analysis of capacity utilization, it is shown that in the case of high capacity utilization of the system, it can provide performance with obvious advantages. In the case of low system capacity utilization, it can still provide performance that is better than or at least not worse than other common dispatch methods. This study also proposes the future research direction. This study can be used as the research basis and basis for the optimization of operations dispatch and preventive maintenance of more complex production systems in the future. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/83511 |
DOI: | 10.6342/NTU202203096 |
Fulltext Rights: | 未授權 |
Appears in Collections: | 工業工程學研究所 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
U0001-0209202211124400.pdf Restricted Access | 1.96 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.