.png){kind=icon}
.png){kind=icon}
Return
Environmentally Aware Quantum-Inspired Reinforcement Learning Framework for Green Power and Resource Optimization in 6G Heterogeneous Networks
.png){kind=icon}
OA
AI
S
K
DOI：10.1109/tgcn.2026.3708287.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The sixth-generation (6G) heterogeneous networks require an intelligent and sustainable resource management system to support ultra-dense connectivity, changing traffic patterns, and energy-efficient operations. In existence, the conventional reinforcement learning (RL) methods lack operation due to slow convergence, inefficient exploration, and limited adaptability under large-scale 6G environments with stochastic energy-harvesting (EH) conditions and varying carbon-intensity levels. After studying to overcome these issues, this paper proposes a hybrid quantum-inspired reinforcement learning with grey wolf optimization and Grover search (QIRL-GWO-Grover) framework for the environmentally aware resource allocation in 6G HetNets. In the proposed framework, GWO-assisted policy initialization is utilized for generating high-quality candidate actions. Grover-inspired amplitude amplification is used to accelerate exploration by amplifying the selection probability of high-reward actions. Besides, the quantum-inspired probabilistic action representation allows adaptive power allocation and resource-block scheduling under dynamic network conditions. Environmental factors including EH dynamics, grid carbon intensity, interference variation, and user QoS requirements are incorporated directly into the state and reward design. Simulation results demonstrate that the proposed framework consistently outperforms DQN, DDPG, Q-learning, Federated RL (FRL), Multi-Agent RL (MARL), and greedy allocation schemes in terms of energy efficiency, throughput, delay, carbon-emission reduction, and convergence speed. The proposed QIRL–GWO–Grover framework achieves up to a 32% improvement in energy efficiency and 28% reduction in carbon emissions while maintaining stable convergence under ultra-dense and dynamic 6G conditions. These results demonstrate the effectiveness of the proposed framework for sustainable and intelligent green resource optimization in future 6G wireless systems.
Keywords:
Quantum-inspired reinforcement learning
6G heterogeneous networks
energy harvesting small cells
green resource allocation
Journal
I
IF:
6.7
Papers:
1.3K
Citations:
4.3K
Organization
Cited Papers