Allocation de ressources dans les réseaux cellulaires des nouvelles générations de la téléphonie mobile par intelligence artificielle.

Abstract

The significant advancements in mobile applications and the growing demand from users have highlighted the need for optimal design solutions in 5G networks. Resource allocation is a key issue for these networks, where a base station allocates resources to a set of geographically dispersed users in a cell to achieve a global .Optimal resource allocation is crucial to maximize the performance of wireless networks in terms of throughput, latency, spectral efficiency, and energy efficiency. This goal can only be achieved by improving traditional networking techniques through artificial intelligence. In this context, this thesis proposes by adopting a hybrid approach. We present a new resource allocation technique that integrates a radial basis function neural network (RBFNN) with the Archimedes optimization algorithm (AOA). The RBFNN is designed to and anticipate user needs, in order to reallocate resources efficiently. Simulation results show that resource allocation using the AOA-RBFNN algorithm provides efficient throughput distribution with reduced latency. Additionally, it allows for predicting user demands while improving the use of available resources and avoiding resource wastage.