Volume 55, 2021Regular articles published in advance of the transition of the journal to Subscribe to Open (S2O). Free supplement sponsored by the Fonds National pour la Science Ouverte
|Page(s)||S1339 - S1354|
|Published online||02 March 2021|
Dynamic vs. static maintenance rate policies for multi-state queueing systems
Business School, Sun Yat-sen University, Guangzhou 510275, P.R. China
2 School of Mathematics and Statistics, Central South University, Changsha 410083, P.R. China
3 School of Science, Wuhan University of Technology, Wuhan 430070, P.R. China
* Corresponding author: email@example.com
Accepted: 3 April 2020
This paper considers a single-server queueing system with server breakdowns. When the server fails, it is sent to repair immediately and its maintenance rate is allowed to be adjustable. The goal is to find the optimal maintenance rates that minimize the long-run average cost of the system. We address the static and dynamic maintenance control problems respectively. For the static control model, we derive the stationary system performances and the explicit solution of the optimal maintenance rate. Then regarding the dynamic maintenance problem, we formulate it as a Markov decision process (MDP) and the optimal dynamic policy is proved to be a threshold policy. Based on the structure of optimal dynamic policy, we construct a performance evaluation for computing efficiently the optimal threshold and average cost. Finally, a comparative study of the dynamic and static policies is presented by numerical experiment which shows the impact of system parameters on the optimal maintenance policies.
Mathematics Subject Classification: 90B22 / 60K25
Key words: Queueing system / maintenance rate / MDP / performance evaluation
© EDP Sciences, ROADEF, SMAI 2021
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.