Abstract—Millimeter-wave and terahertz systems require adaptive interference-aware beamforming under stringent hardware and channel constraints. Existing digital and codebook-based analog schemes either depend on accurate channel state information or fail to handle interference effectively. This work proposes a novel hybrid learning– optimization framework that, for the first time, integrates Deep Reinforcement Learning (DRL) with the Grey Wolf Optimizer (GWO) to enable Channel State Information (CSI)-free beamforming. The DRL agent learns beamforming policies directly from received power feedback, while GWO performs metaheuristic refinement guided by a signal-to-interference-plus-noise ratio–based fitness function. The pretrained DRL policy accelerates convergence and enhances initialization quality. Numerical evaluations show consistent superiority over both GWO-only and DRL–particle swarm optimization counterparts, achieving about 29 dB steady-state performance within 500 ms, sub-130 ms online inference across quantization resolutions up to 5 bits, and near-linear scalability with antenna size. These results confirm the proposed hybrid DRL–GWO framework as a scalable, efficient, and CSI-free solution for real-time interference suppression in next-generation millimeter-wave and terahertz networks.

Keywords—interference suppression, deep reinforcement learning, metaheuristic optimization, grey wolf optimizer, channel state information free design, wireless communication networks


Cite: L. T. Trang, N. V. Cuong, and T. V. Luyen, “Hybrid Deep Reinforcement Learning and Grey Wolf Optimizer for Interference-Aware Beamforming in Wireless Networks," Journal of Communications, vol. 21, no. 1, pp. 57-69, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).