Premium
A docitive Q‐learning approach towards joint resource allocation and power control in self‐organised femtocell networks
Author(s) -
Shahid Adnan,
Aslam Saleem,
Kim Hyung Seok,
Lee KyungGeun
Publication year - 2015
Publication title -
transactions on emerging telecommunications technologies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.366
H-Index - 47
ISSN - 2161-3915
DOI - 10.1002/ett.2800
Subject(s) - femtocell , computer science , resource allocation , quality of service , power control , throughput , control (management) , power (physics) , computer network , wireless , artificial intelligence , base station , telecommunications , physics , quantum mechanics
Abstract Femtocell is a technology that contributes towards the escalation of coverage as well as throughput. By virtue of uncertain deployment of femtocells, self‐organisation is a viable solution for resource allocation. In this study, we are projecting a docitive Q‐learning (DQL) paradigm for joint resource allocation and power control (JRAPC). Moreover, the proposed learning paradigm is compared with independent Q‐learning for the same JRAPC problem. In the proposed DQL paradigm, femto base stations, which are agents, learn the strategies by exploiting Q‐learning and share their learned strategies with their neighbours. Concerning the shared channel environment, the problem function is formulated as the maximisation of femtocell capacity while maintaining the quality of service requirement of the macrousers. The impact of the proposed DQL paradigm is investigated on system capacity and femtocell capacity. Furthermore, comparison is carried out with the considered independent learning paradigm in terms of convergence, min‐max capacity and the effect of femtocell density. Also, the fairness index is computed to have further insight. The results illustrate that DQL‐based JRAPC outperforms its counterpart. Copyright © 2014 John Wiley & Sons, Ltd.