| 講演抄録/キーワード |
| 講演名 |
2014-03-03 14:50
Cell Selection in a Dynamic Femtocell Environment: Restless Multi-Armed Bandit Formulation ○Chaima Dhahri・Tomoaki Ohtsuki(Keio Univ.) RCS2013-327 |
| 抄録 |
(和) |
(まだ登録されていません) |
| (英) |
In this report, we model the problem of cell selection in open-access femtocell networks as a decentralized restless multi-armed bandit (RMAB) with unknown dynamics and multiple players. This is a classic exploration/exploitation problem where each user is trying to learn the best channel to maximize its capacity and reduce its frequency of handover. Each channel is modeled as finite state Markov chain with a Markovian reward. To solve this problem, we refer to the decentralized Restless Upper Confidence Bound (RUCB) algorithm that achieves a logarithmic regret over time for MAB problem. Also, we extend this algorithm to handle dynamic environment by applying a change point detection test based on Page-Hinkley test (PHT). Simulation results show that
? our two proposals come very close to the performance of opportunistic method in terms of capacity, while fewer average number of handovers are required.
? the use of a change point test allows better performance in terms of capacity particularly in a dynamic environment. |
| キーワード |
(和) |
/ / / / / / / |
| (英) |
cell selection / femtocell / multi-armed bandit / dynamic environment / / / / |
| 文献情報 |
信学技報, vol. 113, no. 456, RCS2013-327, pp. 127-131, 2014年3月. |
| 資料番号 |
RCS2013-327 |
| 発行日 |
2014-02-24 (RCS) |
| ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| PDFダウンロード |
RCS2013-327 |