University of Twente Student Theses

Login
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.

Learned Communication for Multi-Agent Spectrum Allocation in D2D Underlay Networks

Neija, N.R. (2025) Learned Communication for Multi-Agent Spectrum Allocation in D2D Underlay Networks.

[img] PDF
4MB
Abstract:In increasingly dense wireless networks efficient and adaptive spectrum allocation is essential for maintaining performance. Reinforcement learning (RL) offers a promising framework for decentralised decision making, and Q-learning has been successfully applied to resource allocation tasks in underlay device-to-device (D2D) communication. Further improvements may be possible by enabling coordination among agents through learned communication. This paper investigates the integration of learned inter-agent communication, using the CommNet architecture, into a multi-agent RL framework for D2D spectrum allocation. The proposed model is evaluated in a two-tier heterogeneous network environment with stochastic user and base station distributions. Simulation results show that, although the performance gains are modest, the CommNet-based model consistently outperforms the Q-learning baseline across a range of user densities. It achieves higher D2D throughput and lower outage probabilities for cellular users. The introduced communication overhead, while manageable in theory, requires infrastructure to facilitate a shared message channel, possibly limiting applicability in practical deployments.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:https://purl.utwente.nl/essays/107393
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page