講演抄録/キーワード |
講演名 |
2023-03-01 16:40
Study of Deep Reinforcement Learning for Wireless Multihop Networks ○Cui Zhihan・Khun Aung thura phyo・Lim Yuto・Tan Yasuo(JAIST) SeMI2022-113 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
In beyond 5G network, the device-to-device communications has been actively studied. These devices are wirelessly connected to each other and can receive and send information by forming a wireless multihop network (WMN). In WMN, there are still some issues that need to be resolved, like due to the uncertainty of source node choosing the path to send the message, the performance of network capacity can degrade drastically. Also, the high interference between nodes limits the transmission rate of the path. To solve these problems, in this research we propose two novel path selection algorithms called SNR-based learning path selection (NLPS) algorithm and SINR-based learning path selection (INLPS) algorithm based on deep reinforcement learning (DRL) to select best multihop path from source node to destination node with highest transmission rate and lowest interference and use factor graph (FG) representation to reduce the heavy iteration. Nested lattice code (NLC) is used in compute-and-forward strategy to reduce the time slots. Our simulation results reveal that NLPS use less iterations while INLPS can achieve higher network capacity. With NLC, the network capacity increases more. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Wireless Multihop Networks / Deep Reinforcement Learning / Factor Graph / Nested Lattice Code / Network Capacity / Computation Time / / |
文献情報 |
信学技報, vol. 122, no. 390, SeMI2022-113, pp. 37-42, 2023年2月. |
資料番号 |
SeMI2022-113 |
発行日 |
2023-02-21 (SeMI) |
ISSN |
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 |
SeMI2022-113 |
|