ScholarWorks@성균관대학교

초록

Wireless resource allocation is a fundamental challenge in modern wireless networks, requiring efficient allocation of limited radio resources while satisfying strict quality of service (QoS) constraints. Traditional optimization techniques, including heuristic-based approaches, struggle with the scalability and computational complexity of large-scale wireless resource allocation problems. In this work, we formulate the wireless resource allocation problem as a combinatorial optimization (CO) problem and leverage a deep reinforcement learning (DRL) framework to efficiently solve it. Unlike conventional DRL methods that optimize decision-making over time domain, our algorithmic-step DRL iteratively allocates radio resources to maximize future rewards. Through extensive simulations, we compare our approach against heuristic baselines, demonstrating better computational efficiency and performance quality. Our results highlight the potential of DRL as a scalable and effective alternative for wireless resource allocation in next-generation networks.

키워드