Towards 5G: a reinforcement learning-based scheduling solution for data traffic management

2.50
Hdl Handle:
http://hdl.handle.net/10547/622852
Title:
Towards 5G: a reinforcement learning-based scheduling solution for data traffic management
Authors:
Comşa, Ioan-Sorin ( 0000-0002-9121-0286 ) ; Zhang, Sijing; Aydin, Mehmet Emin ( 0000-0002-4890-5648 ) ; Kuonen, Pierre; Lu, Yao ( 0000-0001-6241-8718 ) ; Trestian, Ramona ( 0000-0003-3315-3081 ) ; Ghinea, Gheorghiţă ( 0000-0003-2578-5580 )
Abstract:
Dominated by delay-sensitive and massive data applications, radio resource management in 5G access networks is expected to satisfy very stringent delay and packet loss requirements. In this context, the packet scheduler plays a central role by allocating user data packets in the frequency domain at each predefined time interval. Standard scheduling rules are known limited in satisfying higher Quality of Service (QoS) demands when facing unpredictable network conditions and dynamic traffic circumstances. This paper proposes an innovative scheduling framework able to select different scheduling rules according to instantaneous scheduler states in order to minimize the packet delays and packet drop rates for strict QoS requirements applications. To deal with real-time scheduling, the Reinforcement Learning (RL) principles are used to map the scheduling rules to each state and to learn when to apply each. Additionally, neural networks are used as function approximation to cope with the RL complexity and very large representations of the scheduler state space. Simulation results demonstrate that the proposed framework outperforms the conventional scheduling strategies in terms of delay and packet drop rate requirements.
Affiliation:
Brunel University; University of Bedfordshire; University of the West of England; University of Applied Sciences of Western Switzerland; University of Fribourg; Middlesex University
Citation:
Comsa I, Zhang S, Aydin M, Kuonen P, Lu Y, Trestian R, Ghinea G (2018) 'Towards 5G: a reinforcement learning-based scheduling solution for data traffic management', IEEE Transactions on Network and Service Management, (), pp.-.
Publisher:
IEEE
Journal:
IEEE Transactions on Network and Service Management
Issue Date:
6-Aug-2018
URI:
http://hdl.handle.net/10547/622852
DOI:
10.1109/TNSM.2018.2863563
Additional Links:
https://ieeexplore.ieee.org/document/8425580/
Type:
Article
Language:
en
ISSN:
1932-4537
EISSN:
1932-4537
Appears in Collections:
Computing

Full metadata record

DC FieldValue Language
dc.contributor.authorComşa, Ioan-Sorinen
dc.contributor.authorZhang, Sijingen
dc.contributor.authorAydin, Mehmet Eminen
dc.contributor.authorKuonen, Pierreen
dc.contributor.authorLu, Yaoen
dc.contributor.authorTrestian, Ramonaen
dc.contributor.authorGhinea, Gheorghiţăen
dc.date.accessioned2018-09-13T09:57:49Z-
dc.date.available2018-09-13T09:57:49Z-
dc.date.issued2018-08-06-
dc.identifier.citationComsa I, Zhang S, Aydin M, Kuonen P, Lu Y, Trestian R, Ghinea G (2018) 'Towards 5G: a reinforcement learning-based scheduling solution for data traffic management', IEEE Transactions on Network and Service Management, (), pp.-.en
dc.identifier.issn1932-4537-
dc.identifier.doi10.1109/TNSM.2018.2863563-
dc.identifier.urihttp://hdl.handle.net/10547/622852-
dc.description.abstractDominated by delay-sensitive and massive data applications, radio resource management in 5G access networks is expected to satisfy very stringent delay and packet loss requirements. In this context, the packet scheduler plays a central role by allocating user data packets in the frequency domain at each predefined time interval. Standard scheduling rules are known limited in satisfying higher Quality of Service (QoS) demands when facing unpredictable network conditions and dynamic traffic circumstances. This paper proposes an innovative scheduling framework able to select different scheduling rules according to instantaneous scheduler states in order to minimize the packet delays and packet drop rates for strict QoS requirements applications. To deal with real-time scheduling, the Reinforcement Learning (RL) principles are used to map the scheduling rules to each state and to learn when to apply each. Additionally, neural networks are used as function approximation to cope with the RL complexity and very large representations of the scheduler state space. Simulation results demonstrate that the proposed framework outperforms the conventional scheduling strategies in terms of delay and packet drop rate requirements.en
dc.language.isoenen
dc.publisherIEEEen
dc.relation.urlhttps://ieeexplore.ieee.org/document/8425580/en
dc.rightsGreen - can archive pre-print and post-print or publisher's version/PDF-
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/4.0/*
dc.subject5G mobile communicationen
dc.subjectpacket schedulingen
dc.subjectoptimisationen
dc.subjectneural networksen
dc.subjectradio resource managementen
dc.subjectreinforcement Learningen
dc.subjectG420 Networks and Communicationsen
dc.titleTowards 5G: a reinforcement learning-based scheduling solution for data traffic managementen
dc.typeArticleen
dc.identifier.eissn1932-4537-
dc.contributor.departmentBrunel Universityen
dc.contributor.departmentUniversity of Bedfordshireen
dc.contributor.departmentUniversity of the West of Englanden
dc.contributor.departmentUniversity of Applied Sciences of Western Switzerlanden
dc.contributor.departmentUniversity of Fribourgen
dc.contributor.departmentMiddlesex Universityen
dc.identifier.journalIEEE Transactions on Network and Service Managementen
dc.date.updated2018-09-13T09:49:36Z-
This item is licensed under a Creative Commons License
Creative Commons
All Items in UOBREP are protected by copyright, with all rights reserved, unless otherwise indicated.