Distributed Learning for Channel Allocation over a Shared Spectrum

S. M. Zafaruddin; Ilai Bistritz; Amir Leshem; Dusit Niyato

doi:10.1109/JSAC.2019.2933966

Distributed Learning for Channel Allocation over a Shared Spectrum

S. M. Zafaruddin^*, Ilai Bistritz, Amir Leshem, Dusit Niyato

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

24 Scopus citations

Abstract

Channel allocation is the task of assigning channels to users such that some objective (e.g., sum-rate) is maximized. In centralized networks such as cellular networks, this task is carried by the base station (BS) which gathers the channel state information (CSI) from the users and computes the optimal solution. In distributed networks such as ad-hoc and device-to-device (D2D) networks, no BS exists and conveying global CSI between users is costly or simply impractical. When the CSI is time varying and unknown to the users, the users face the challenge of both learning the channel statistics online and converging to a good channel allocation. This introduces a multi-armed bandit (MAB) scenario with multiple decision makers. If two or more users choose the same channel, a collision occurs and they all receive zero reward. We propose a distributed channel allocation algorithm that each user runs and converges to the optimal allocation while achieving an order optimal regret of O log T , where T denotes the length of time horizon. The algorithm is based on a carrier sensing multiple access (CSMA) implementation of the distributed auction algorithm. It does not require any exchange of information between users. Users need only to observe a single channel at a time and sense if there is a transmission on that channel, without decoding the transmissions or identifying the transmitting users. We demonstrate the performance of our algorithm using simulated LTE and 5G channels.

Original language	American English
Article number	8792155
Pages (from-to)	2337-2349
Number of pages	13
Journal	IEEE Journal on Selected Areas in Communications
Volume	37
Issue number	10
DOIs	https://doi.org/10.1109/JSAC.2019.2933966
State	Published - Oct 2019
Externally published	Yes

Bibliographical note

Funding Information:
Manuscript received December 15, 2018; revised April 5, 2019; accepted May 20, 2019. Date of publication August 8, 2019; date of current version September 16, 2019. This work was supported in part by the ISF-NRF Joint Research Program under Grant ISF 2277/16 and Grant ISF 1644/18, in part by WASP/NTU M4082187 under Grant 4080, in part by the Singapore MOE Tier 1 under Grant 2017-T1-002-007 RG122/17, and in part by NRF2015-NRF-ISF001-2277. This article was presented at the 20th IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Cannes, France, July 2019. The work of S. M. Zafaruddin was supported in part by the Israeli Planning and Budget Committee (PBC) Post-Doctoral Fellowship (2016–2018). (Corresponding author: S. M. Zafaruddin.) S. M. Zafaruddin was with the Faculty of Engineering, Bar-Ilan University, Ramat Gan 5290002, Israel. He is now with the Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science at Pilani, Pilani 333031, India (e-mail: syed.zafaruddin@pilani.bits-pilani.ac.in).

Publisher Copyright:
© 1983-2012 IEEE.

Keywords

Distributed channel allocation
dynamic spectrum accesses
multiplayer multi-armed bandit
online learning
resource management

Access to Document

10.1109/JSAC.2019.2933966

Cite this

@article{915cd4d918a24930b767d3a4b1749917,

title = "Distributed Learning for Channel Allocation over a Shared Spectrum",

abstract = "Channel allocation is the task of assigning channels to users such that some objective (e.g., sum-rate) is maximized. In centralized networks such as cellular networks, this task is carried by the base station (BS) which gathers the channel state information (CSI) from the users and computes the optimal solution. In distributed networks such as ad-hoc and device-to-device (D2D) networks, no BS exists and conveying global CSI between users is costly or simply impractical. When the CSI is time varying and unknown to the users, the users face the challenge of both learning the channel statistics online and converging to a good channel allocation. This introduces a multi-armed bandit (MAB) scenario with multiple decision makers. If two or more users choose the same channel, a collision occurs and they all receive zero reward. We propose a distributed channel allocation algorithm that each user runs and converges to the optimal allocation while achieving an order optimal regret of O log T , where T denotes the length of time horizon. The algorithm is based on a carrier sensing multiple access (CSMA) implementation of the distributed auction algorithm. It does not require any exchange of information between users. Users need only to observe a single channel at a time and sense if there is a transmission on that channel, without decoding the transmissions or identifying the transmitting users. We demonstrate the performance of our algorithm using simulated LTE and 5G channels.",

keywords = "Distributed channel allocation, dynamic spectrum accesses, multiplayer multi-armed bandit, online learning, resource management",

author = "Zafaruddin, {S. M.} and Ilai Bistritz and Amir Leshem and Dusit Niyato",

note = "Funding Information: Manuscript received December 15, 2018; revised April 5, 2019; accepted May 20, 2019. Date of publication August 8, 2019; date of current version September 16, 2019. This work was supported in part by the ISF-NRF Joint Research Program under Grant ISF 2277/16 and Grant ISF 1644/18, in part by WASP/NTU M4082187 under Grant 4080, in part by the Singapore MOE Tier 1 under Grant 2017-T1-002-007 RG122/17, and in part by NRF2015-NRF-ISF001-2277. This article was presented at the 20th IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Cannes, France, July 2019. The work of S. M. Zafaruddin was supported in part by the Israeli Planning and Budget Committee (PBC) Post-Doctoral Fellowship (2016–2018). (Corresponding author: S. M. Zafaruddin.) S. M. Zafaruddin was with the Faculty of Engineering, Bar-Ilan University, Ramat Gan 5290002, Israel. He is now with the Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science at Pilani, Pilani 333031, India (e-mail: syed.zafaruddin@pilani.bits-pilani.ac.in). Publisher Copyright: {\textcopyright} 1983-2012 IEEE.",

year = "2019",

month = oct,

doi = "10.1109/JSAC.2019.2933966",

language = "American English",

volume = "37",

pages = "2337--2349",

journal = "IEEE Journal on Selected Areas in Communications",

issn = "0733-8716",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "10",

}

TY - JOUR

T1 - Distributed Learning for Channel Allocation over a Shared Spectrum

AU - Zafaruddin, S. M.

AU - Bistritz, Ilai

AU - Leshem, Amir

AU - Niyato, Dusit

N1 - Funding Information: Manuscript received December 15, 2018; revised April 5, 2019; accepted May 20, 2019. Date of publication August 8, 2019; date of current version September 16, 2019. This work was supported in part by the ISF-NRF Joint Research Program under Grant ISF 2277/16 and Grant ISF 1644/18, in part by WASP/NTU M4082187 under Grant 4080, in part by the Singapore MOE Tier 1 under Grant 2017-T1-002-007 RG122/17, and in part by NRF2015-NRF-ISF001-2277. This article was presented at the 20th IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Cannes, France, July 2019. The work of S. M. Zafaruddin was supported in part by the Israeli Planning and Budget Committee (PBC) Post-Doctoral Fellowship (2016–2018). (Corresponding author: S. M. Zafaruddin.) S. M. Zafaruddin was with the Faculty of Engineering, Bar-Ilan University, Ramat Gan 5290002, Israel. He is now with the Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science at Pilani, Pilani 333031, India (e-mail: syed.zafaruddin@pilani.bits-pilani.ac.in). Publisher Copyright: © 1983-2012 IEEE.

PY - 2019/10

Y1 - 2019/10

N2 - Channel allocation is the task of assigning channels to users such that some objective (e.g., sum-rate) is maximized. In centralized networks such as cellular networks, this task is carried by the base station (BS) which gathers the channel state information (CSI) from the users and computes the optimal solution. In distributed networks such as ad-hoc and device-to-device (D2D) networks, no BS exists and conveying global CSI between users is costly or simply impractical. When the CSI is time varying and unknown to the users, the users face the challenge of both learning the channel statistics online and converging to a good channel allocation. This introduces a multi-armed bandit (MAB) scenario with multiple decision makers. If two or more users choose the same channel, a collision occurs and they all receive zero reward. We propose a distributed channel allocation algorithm that each user runs and converges to the optimal allocation while achieving an order optimal regret of O log T , where T denotes the length of time horizon. The algorithm is based on a carrier sensing multiple access (CSMA) implementation of the distributed auction algorithm. It does not require any exchange of information between users. Users need only to observe a single channel at a time and sense if there is a transmission on that channel, without decoding the transmissions or identifying the transmitting users. We demonstrate the performance of our algorithm using simulated LTE and 5G channels.

AB - Channel allocation is the task of assigning channels to users such that some objective (e.g., sum-rate) is maximized. In centralized networks such as cellular networks, this task is carried by the base station (BS) which gathers the channel state information (CSI) from the users and computes the optimal solution. In distributed networks such as ad-hoc and device-to-device (D2D) networks, no BS exists and conveying global CSI between users is costly or simply impractical. When the CSI is time varying and unknown to the users, the users face the challenge of both learning the channel statistics online and converging to a good channel allocation. This introduces a multi-armed bandit (MAB) scenario with multiple decision makers. If two or more users choose the same channel, a collision occurs and they all receive zero reward. We propose a distributed channel allocation algorithm that each user runs and converges to the optimal allocation while achieving an order optimal regret of O log T , where T denotes the length of time horizon. The algorithm is based on a carrier sensing multiple access (CSMA) implementation of the distributed auction algorithm. It does not require any exchange of information between users. Users need only to observe a single channel at a time and sense if there is a transmission on that channel, without decoding the transmissions or identifying the transmitting users. We demonstrate the performance of our algorithm using simulated LTE and 5G channels.

KW - Distributed channel allocation

KW - dynamic spectrum accesses

KW - multiplayer multi-armed bandit

KW - online learning

KW - resource management

UR - http://www.scopus.com/inward/record.url?scp=85070671287&partnerID=8YFLogxK

U2 - 10.1109/JSAC.2019.2933966

DO - 10.1109/JSAC.2019.2933966

M3 - Article

AN - SCOPUS:85070671287

SN - 0733-8716

VL - 37

SP - 2337

EP - 2349

JO - IEEE Journal on Selected Areas in Communications

JF - IEEE Journal on Selected Areas in Communications

IS - 10

M1 - 8792155

ER -

Distributed Learning for Channel Allocation over a Shared Spectrum

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this