首页 | 本学科首页   官方微博 | 高级检索  
     检索      


Scheduling policies for a repair shop problem
Authors:William Kun Liang  Bar?? Balc?og?lu  Robert Svaluto
Institution:1. Cosma International Group, Magna International, 2550 Steeles Avenue East, Brampton, ON, L6T 5R3, Canada
2. Faculty of Engineering and Natural Sciences, Sabanc? University, Orhanl?-Tuzla, 34956, Istanbul, Turkey
3. IMAX Corporation, 2525 Speakman Drive, Mississauga, Ontario, L5K 1B1, Canada
Abstract:In this paper, we analyze a repair shop serving several fleets of machines that fail from time to time. To reduce downtime costs, a continuous-review spare machine inventory is kept for each fleet. A spare machine, if available on stock, is installed instantaneously in place of a broken machine. When a repaired machine is returned from the repair shop, it is placed in inventory for future use if the fleet has the required number of machines operating. Since the repair shop is shared by different fleets, choosing which type of broken machine to repair is crucial to minimize downtime and holding costs. The optimal policy of this problem is difficult to characterize, and, therefore, is only formulated as a Markov Decision Process to numerically compute the optimal cost and base-stock level for each spare machine inventory. As an alternative, we propose the dynamic Myopic(R) policy, which is easy to implement, yielding costs very close to the optimal. Most of the time it outperforms the static first-come-first-served, and preemptive-resume priority policies. Additionally, via our numerical study, we demonstrate that repair shop pooling is better than reserving a repair shop for each fleet.
Keywords:
本文献已被 SpringerLink 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号