Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 2
Encontrada(s) 1 página(s)

RIUMA. Repositorio Institucional de la Universidad de Málaga
Publicación Artículo científico (JournalArticle)

Latency Fairness Optimization on Wireless Networks Through Deep Reinforcement Learning

  • López-Sánchez, María
  • Villena-Rodríguez, Alejandro
  • Gómez-Paredes, Gerardo
  • Martín-Vega, Francisco-Javier
  • Aguayo-Torres, María del Carmen
https://openpolicyfinder.jisc.ac.uk/id/publication/3582, In this paper, we propose a novel deep reinforcement learning (DRL) framework to maximize user fairness in terms of delay. To this end, we devise a new version of the modified largest weighted delay first (M-LWDF) algorithm, called β-M-LWDF, aiming to fulfill an appropriate balance between user fairness and average delay. This balance is defined as a feasible region on the cumulative distribution function (CDF) of the user delay that allows identifying unfair states, feasible-fair states, and over-fair states. Simulation results reveal that our proposed framework outperforms traditional resource allocation techniques in terms of latency fairness and average delay., Unión Europea, Junta de Andalucía, Universidad de Málaga, Agencia Estatal de Investigación
RIUMA. Repositorio Institucional de la Universidad de Málaga
Publicación Artículo científico (JournalArticle)

A Physical Layer Multicast Precoding and Grouping Scheme for Bandwidth Minimization.

  • Martín-Vega, Francisco Javier
  • Rostami Ghadi, Farshad
  • López-Martínez, Francisco Javier
  • Gómez-Paredes, Gerardo
Physical layer multicasting exploits multiple antennas at the transmitter side to deliver a common message to a group of K users. To this end, two formulations have been well addressed in the literature: i) the max-min-fair criterion, which maximizes the signal-to-noise ratio (SNR) of the worst user for a fixed transmit power; and ii) the quality of service (QoS) formulation, which minimizes the transmit power subject to a target SNR. Nevertheless, it is known that the performance and complexity of these approaches is severely degraded as the group size grows. In this paper, we propose a different formulation that aims at minimizing the required bandwidth needed to provide the multicast service. This is achieved by dividing the users into smaller groups and assigning the bandwidth required to provide a target rate to each group. Contrary to the common belief, it is shown that dividing the users into different groups that use orthogonal bandwidth allocations can lead to a smaller aggregated bandwidth than the single-group with single bandwidth allocation counterpart, if an intelligent grouping scheme is used. An iterative algorithm to derive the optimal number of groups is presented with an stopping criterion to reduce the numerical complexity. It is shown through simulation that our proposed approach greatly reduces the required bandwidth compared to existing schemes that rely on single bandwidth allocation. Interestingly, results reveal that our proposed scheme also leads to a greater SNR for a randomly chosen user, and it reduces the variance of the required bandwidth, which eases the implementation in real networks., Unión Europea, Junta de Andalucía, Universidad de Málaga