Energy-Efficient Dynamic Enhanced Inter-Cell Interference Coordination Scheme Based on Deep Reinforcement Learning in...

Area of Science:

• Wireless Communication
• Network Engineering
• Artificial Intelligence

Background:

• 5G networks offer high speeds and low latency but face challenges in heterogeneous cloud radio access networks (H-CRAN).
• Managing inter-cell interference and conserving energy are critical issues in H-CRAN deployments.
• Conventional methods for interference coordination have limitations in optimizing energy efficiency.

Purpose of the Study:

• To propose a novel energy-efficient, dynamic enhanced inter-cell interference coordination (eICIC) scheme for 5G H-CRAN.
• To integrate energy consumption and quality of service (QoS) into the eICIC optimization process.
• To enhance the sustainability and performance of 5G networks.

Main Methods:

• Developed a deep reinforcement learning (DRL) based eICIC scheme.
• Introduced transmission power during almost blank subframes (TPA) and channel quality indicator (CQI) threshold of victim user equipments (CTV) as optimization parameters.
• Modeled signal-to-interference-plus-noise ratio (SINR) and service rates to define energy-utility efficiency.

Main Results:

• The proposed DRL-based eICIC scheme achieved significant energy savings, up to 70%.
• The scheme demonstrated enhanced Quality of Service (QoS) satisfaction.
• The energy-utility efficiency metric effectively balanced energy savings and QoS.

Conclusions:

• The novel eICIC scheme offers a promising solution for energy-efficient and high-performance 5G H-CRAN.
• DRL is an effective approach for dynamic resource management in complex wireless networks.
• The findings contribute to the development of sustainable and efficient future wireless communication systems.