Projecte llegit
Títol: Fast heuristics for optimising 5G networks
Director/a: ZOLA, ENRICA VALERIA
Departament: ENTEL
Títol: Fast heuristics for optimising 5G networks
Data inici oferta: 20-11-2016 Data finalització oferta: 20-06-2017
Estudis d'assignació del projecte:
MU MASTEAM 2015
| Tipus: Individual | |
| Lloc de realització: EETAC | |
| Paraules clau: | |
| Fast heuristics, 5G Optimisation, System performance | |
| Descripció del contingut i pla d'activitats: | |
| A dense deployment of small cells is one of the key characteristics envisioned for future
5G mobile networks in order to provide the required capacity increase where needed. In this context, it becomes very important to develop strategies to turn on and off the small cells according to the users needs, so to reduce costs and CO2 emissions. This work focuses on developing heuristics, which help minimizing the total power consumption of 5G HetNets while providing the required capacity and coverage. The second objective of this work is to analyse the heuristics in several scenarios in order to assess their validity and to estimate the trade-off between the solution obtained and the true optimum. |
|
| Overview (resum en anglès): | |
| This study focuses on 5G network, which deploys small cells to form multi-hop topologies using high capacity backhaul wireless links to provide localized capacity. Nowadays, high energy efficiency is very important because powering on unnecessarily a massive amount of macro cells or small cells may lead to increased expenses, CO2 emission and environmental destruction. Based on a given MILP that solves the energy consumption optimization problem in a 5G network, this research proposes a heuristic algorithm based on integer relaxation that accelerates the resolution of the MILP. The heuristic algorithm could diminish the route options by striking out the impossible links or links with lower possibility to be used. Our numerical evaluations demonstrate that the proposed algorithm can find very good solutions in short time and has similar performance in terms of energy efficiency over a large number of traffic scenarios. | |