| Titre : | Exploiting heterogeneity in the development of the next generation 5G mobile networks | | Type de document : | thèse | | Auteurs : | Walid Abdellaziz, Auteur | | Année de publication : | 2016 | | Langues : | Anglais (eng) | | Catégories : | Informatique
| | Mots-clés : | Fifth generation (5G), Heterogeneous networks, Energy Efficiency, Load balancing,
Handover, LTE/LTE-A, Multi-RAT, HetNets, Small cells, Small base Stations (SBSs), Decoupling Uplink
Downlink (DUD), Multi-homing. | | Index. décimale : | Doct/265 | | Résumé : | Exploiting the dense heterogeneity in terms of multiple radio access technologies (LTE/LTEA, WiMax,
HSDPA, Wi-Fi, ...etc.), as well as different cell sizes (macro, micro, metro, pico, and femto cells) is a
promising way to achieve the main goals of the fifth generation (5G) of cellular mobile networks. However,
the deployment of new network architectures based on this dense heterogeneity faces many big technical
challenges to be overcome. This PhD thesis addresses some major technical challenges related to vertical
handoff management, interference management, QoS management, backhauling, energy efficiency, as well
as load balancing that need huge efforts for the deployment of 5G networks in the near future of 2020.
In this PhD thesis, we address some major technical challenges linked to the development of 5G
heterogeneous networks. Furthermore, we propose four contributions based on recent mathematical tools,
methodologies and technologies that are seen by both industrial and academic research communities as
golden key solutions for the success of the deployment of 5G heterogeneous networks.
In the first contribution, we address the problem of network congestion in 5G multi-RAT heterogeneous
networks. This problem eventually occurred in group mobility scenarios due to the vertical handoff decisions
done nearly at the same time by a number of mobile users moving together inside a car or a bus not equipped
with a mobile base station. Therefore, to resolve the problem of network congestion in this situation known
as a group vertical handoff (GVHO), we model our problem as a symmetric congestion game using game
theory. Then, we propose two fully decentralized learning algorithms to reach the nash equilibrium that
represents a fair and efficient solution ensuring load balancing in 5G multi-RAT heterogeneous networks. In
addition, to adapt our solution for high mobility scenarios, we propose a heuristic method dubbed DSSSA
(Decreasing Step Size-Simulated Annealing) incorporated in a hybrid reinforcement learning algorithm to
speed up the convergence time to the nash equilibrium solution.
In the second contribution we deal with the challenges related to downlink interference management,
downlink QoS management and backhauling in the hyper-dense LTE HetNets architectures. Then, we
provide a solution based on a self-organizing network (SON) algorithm combined with the multi-homing
capabilities of macro cellular users, to improve the overall downlink throughput system, as well as to satisfy
the QoS throughput requirements of both home and macro cellular users. Moreover, our solution permits to
reduce the huge overhead signaling induced by centralized scheme solutions in hyper-dense HetNets. Instead
of the downlink channel studied in the previous work, in the contributions 3 and 4, we focus our study on the
uplink channel due to the assumed decoupling uplink downlink (DUD) access. In the third contribution we
propose the same solution as in the second contribution for the uplink channel to improve the energy
efficiency of cellular mobile users. However, due to the technical limitations to deploy many homogeneous
LTE interfaces in mobile devices, in the fourth contribution we consider mobile devices with single LTE
interface, then we formulate the same problem as a many to one matching game using matching theory,
where the two sided of players are respectively macro indoor cellular users and small base stations (SBSs).
Then, based on the preference list of players, we provide the deferred acceptance algorithm to reach the
optimal stable matching consisting of assigning each macro indoor user to the most suitable SBS and vice
versa.
Finally, the proposed solutions are evaluated through extensive numerical simulations and the numerical
results are presented to provide a comparison with the related works found in the literature.
|
Exploiting heterogeneity in the development of the next generation 5G mobile networks [thèse] / Walid Abdellaziz, Auteur . - 2016. Langues : Anglais ( eng) | Catégories : | Informatique
| | Mots-clés : | Fifth generation (5G), Heterogeneous networks, Energy Efficiency, Load balancing,
Handover, LTE/LTE-A, Multi-RAT, HetNets, Small cells, Small base Stations (SBSs), Decoupling Uplink
Downlink (DUD), Multi-homing. | | Index. décimale : | Doct/265 | | Résumé : | Exploiting the dense heterogeneity in terms of multiple radio access technologies (LTE/LTEA, WiMax,
HSDPA, Wi-Fi, ...etc.), as well as different cell sizes (macro, micro, metro, pico, and femto cells) is a
promising way to achieve the main goals of the fifth generation (5G) of cellular mobile networks. However,
the deployment of new network architectures based on this dense heterogeneity faces many big technical
challenges to be overcome. This PhD thesis addresses some major technical challenges related to vertical
handoff management, interference management, QoS management, backhauling, energy efficiency, as well
as load balancing that need huge efforts for the deployment of 5G networks in the near future of 2020.
In this PhD thesis, we address some major technical challenges linked to the development of 5G
heterogeneous networks. Furthermore, we propose four contributions based on recent mathematical tools,
methodologies and technologies that are seen by both industrial and academic research communities as
golden key solutions for the success of the deployment of 5G heterogeneous networks.
In the first contribution, we address the problem of network congestion in 5G multi-RAT heterogeneous
networks. This problem eventually occurred in group mobility scenarios due to the vertical handoff decisions
done nearly at the same time by a number of mobile users moving together inside a car or a bus not equipped
with a mobile base station. Therefore, to resolve the problem of network congestion in this situation known
as a group vertical handoff (GVHO), we model our problem as a symmetric congestion game using game
theory. Then, we propose two fully decentralized learning algorithms to reach the nash equilibrium that
represents a fair and efficient solution ensuring load balancing in 5G multi-RAT heterogeneous networks. In
addition, to adapt our solution for high mobility scenarios, we propose a heuristic method dubbed DSSSA
(Decreasing Step Size-Simulated Annealing) incorporated in a hybrid reinforcement learning algorithm to
speed up the convergence time to the nash equilibrium solution.
In the second contribution we deal with the challenges related to downlink interference management,
downlink QoS management and backhauling in the hyper-dense LTE HetNets architectures. Then, we
provide a solution based on a self-organizing network (SON) algorithm combined with the multi-homing
capabilities of macro cellular users, to improve the overall downlink throughput system, as well as to satisfy
the QoS throughput requirements of both home and macro cellular users. Moreover, our solution permits to
reduce the huge overhead signaling induced by centralized scheme solutions in hyper-dense HetNets. Instead
of the downlink channel studied in the previous work, in the contributions 3 and 4, we focus our study on the
uplink channel due to the assumed decoupling uplink downlink (DUD) access. In the third contribution we
propose the same solution as in the second contribution for the uplink channel to improve the energy
efficiency of cellular mobile users. However, due to the technical limitations to deploy many homogeneous
LTE interfaces in mobile devices, in the fourth contribution we consider mobile devices with single LTE
interface, then we formulate the same problem as a many to one matching game using matching theory,
where the two sided of players are respectively macro indoor cellular users and small base stations (SBSs).
Then, based on the preference list of players, we provide the deferred acceptance algorithm to reach the
optimal stable matching consisting of assigning each macro indoor user to the most suitable SBS and vice
versa.
Finally, the proposed solutions are evaluated through extensive numerical simulations and the numerical
results are presented to provide a comparison with the related works found in the literature.
|
|
Doc/113 
Faire une suggestion Affiner la rechercheExploiting heterogeneity in the development of the next generation 5G mobile networks / Walid Abdellaziz
