利用馬可夫決策過程製定狀態檢修策略;Purposing a Condition Based Maintenance Policy Using Markov Decision Process

NCUIR > School of Management at National Central University > Graduate Institute of Industrial Management > Electronic Thesis & Dissertation > Item 987654321/91809

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/91809

Title:	利用馬可夫決策過程製定狀態檢修策略;Purposing a Condition Based Maintenance Policy Using Markov Decision Process
Authors:	林宏軒;Lin, Hung-Hsuan
Contributors:	工業管理研究所
Keywords:	狀態檢修;Cox 比例風險模型;馬可夫決策過程;Condition Based Maintenance;Cox Proportional Hazard Model;Markov Decision Process
Date:	2023-07-10
Issue Date:	2024-09-19 14:14:32 (UTC+8)
Publisher:	國立中央大學
Abstract:	在現今自動化、大量生產的製造業中，機台維修的議題持續受到人們的關注。維修策略的制定對於管理者來說相當困難，關於機台的信息充滿著不確定性，如運作的時間長短、工作條件的設定、機台效能的衰退，管理者需要在有限資訊的情況下進行決策。在機台維修的研究中，經常利用馬可夫決策過程為模型探討維修時間的規劃，在固定的時間周期進行維修動作的決策，以動態規劃求解得到長期最佳的維修時間計畫，並稱之為最佳決策。我們希望在馬可夫決策過程中考量機台因為工作條件、環境因素不同而有不同的衰退速率或轉移機率。我們希望管理者可以根據機台的資訊即時的估計衰退速率、轉移矩陣，有效的衡量機台所需承擔的風險，可以制定更好的維修計畫使的總成本降低。在此研究中我們假設機台的狀態有兩種，而狀態的衰退速率會依據當前的工作條件有所差異。我們希望利用 Cox 比例風險模型透過機台的工作條件、環境資訊估計各個狀態的失效率，並利用其結果估計馬可夫決策過程中的轉移機率。藉由機台資料即時的更新調整維修計畫，降低維修、營運的成本。;In today’s automated and mass-produced manufacturing industry, the issue of machine maintenance continues to attract people’s attention. It is very difficult for managers to formulate maintenance strategies. The information about machines is diversity. In the research of machine maintenance, the Markov Decision Process is often used as a model to discuss the planning of maintenance time, the decision-making of maintenance actions is made in a fixed time period, and the long-term optimal maintenance solution is obtained by dynamic programming. We want that managers can estimate the deterioration rate and transition probability in real time based on the information of the machine. We want to use the Cox proportional hazard model to estimate the failure rate of the machine through the working conditions and environmental information. Using the results to estimate the transition probability in the Markov Decision Process, and determine the maintenance policy. By updating information to adjust the maintenance plan in real time, the cost of maintenance and operation can be reduced.
Appears in Collections:	[Graduate Institute of Industrial Management] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	16	View/Open

社群 sharing

Loading...