Mission-critical spectrum sharing with decentralized Multi-Agent Reinforcement Learning
Abstract
Motivated by emerging mission-critical applications and an increasingly congested spectrum, we develop a decentralized multi-agent reinforcement learning (MARL) model for dynamic spectrum access. The model enables secondary users to learn effective transmission strategies across shared frequency bands while minimizing collisions with high-priority primary users and among themselves. We design the agent-level learners following a Markov potential game approach, connecting independent local updates to system-level improvement. We instantiate this design using lightweight linear actor-critic learners suitable for resource-constrained edge devices, rather than computationally intensive centralized or deep multi-agent architectures. Across spectrum environments with different incumbent activities, the learned policies adapt their transmission policy and waiting behavior to preserve throughput while greatly reducing transmission collisions relative to random and forecast-aware heuristic baselines. The results establish the value of decentralized MARL and shows up to 96.8% reduction in overall collisions.
Related
- Decentralized Strategies for Finite Population LQG Social Control: A Reinforcement Learning Approach
- Risk-Sensitive Mission Planning for Multi-Target On-Orbit Inspection via Distributional Reinforcement Learning
- Decomposing URLLC Deadline Failures in Beam-Managed 6G Networks: Implications for Mission-Critical Wireless Systems
- Toward Mission-Critical ISAC: Reliable Energy-Aware Coordination in UAV Swarms
- Decentralized Bandit Learning of Markov Perfect Equilibria
BibTeX
@misc{pylorof2026missioncritical,
title = {{Mission-critical spectrum sharing with decentralized Multi-Agent Reinforcement Learning}},
author = {Dimitrios Pylorof and Imtiaz Nasim and Humberto E. Garcia and Vivek Agarwal and Jasni A. Mannil and Mingyue Ji},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.09213},
eprint = {2610.09213},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.09213},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Mission-critical spectrum sharing with decentralized Multi-Agent Reinforcement Learning
AU - Dimitrios Pylorof
AU - Imtiaz Nasim
AU - Humberto E. Garcia
AU - Vivek Agarwal
AU - Jasni A. Mannil
AU - Mingyue Ji
PY - 2026
DA - 2026/10/06
PB - arXiv
DO - 10.48550/arxiv.2610.09213
UR - https://arxiv.org/abs/2610.09213
AB - Motivated by emerging mission-critical applications and an increasingly congested spectrum, we develop a decentralized multi-agent reinforcement learning (MARL) model for dynamic spectrum access. The model enables secondary users to learn effective transmission strategies across shared frequency bands while minimizing collisions with high-priority primary users and among themselves. We design the agent-level learners following a Markov potential game approach, connecting independent local updates to system-level improvement. We instantiate this design using lightweight linear actor-critic learners suitable for resource-constrained edge devices, rather than computationally intensive centralized or deep multi-agent architectures. Across spectrum environments with different incumbent activities, the learned policies adapt their transmission policy and waiting behavior to preserve throughput while greatly reducing transmission collisions relative to random and forecast-aware heuristic baselines. The results establish the value of decentralized MARL and shows up to 96.8% reduction in overall collisions.
ER -