ORIGINAL RESEARCH article
Front. Commun. Netw.
Sec. Wireless Communications
Volume 6 - 2025 | doi: 10.3389/frcmn.2025.1657288
This article is part of the Research TopicMachine Learning-Based Spectrum Occupancy Prediction and Resource Allocation/Management for Wireless Communication SystemsView all 3 articles
Graph-Theoretic Approach to Mobility-Aware Frequency Assignment via Deep Q-Learning
Provisionally accepted
- 1Yonsei University, Seoul, Republic of Korea
- 2Hankuk University of Foreign Studies, Dongdaemun-gu, Republic of Korea
- 3Hanwha Systems Co Ltd, Jung-gu, Republic of Korea
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Due to continuously increasing communication demands, efficient frequency assignment becomes essential. Unlike conventional cellular communication systems, where frequency allocation is managed centrally by a base station, frequency assignment in device-to-device (D2D) communication, particularly in mission-critical scenarios, presents greater complexity. In this study, we model the D2D network as a graph and formulate the frequency assignment problem as a graph coloring problem. While prior research has employed heuristic methods or AI techniques for ordering the nodes to assign frequency, we introduce advanced algorithms and machine learning approach to enhance efficiency. Additionally, we applied net filter discrimination (NFD) to mitigate interference and enable interference-free communication. While previous AI methods focused only on the optimization objective of minimizing the number of assigned frequency blocks, we also focused on the total frequency span. To improve performance, machine learning method is used for the AI-based approach and compared its effectiveness against a baseline greedy algorithm. Through simulation of a randomly generated communication system, we evaluate the proposed methods and demonstrate that our deep Q-learning method significantly enhance frequency assignment efficiency, particularly in complex scenarios. In addition, the mobility for some devices was added to show the effectiveness of the approach we suggested. We anticipate that this approach will contribute to the realization of reliable frequency assignment strategies suitable for mission-critical applications.
Keywords: Graph coloring problem (GCP), Frequency assignment problem (FAP), Greedy algorithm, deep Q-learning (DQN), net filterdiscrimination (NFD), minimum order (MO), minimum span (MS)
Received: 01 Jul 2025; Accepted: 12 Aug 2025.
Copyright: © 2025 Kim, Jeon, Ji, Park and Chae. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Chan-Byoung Chae, Yonsei University, Seoul, Republic of Korea
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.