Survey of Reinforcement-Learning-Based MAC Protocols for Wireless Ad Hoc Networks with a MAC Reference Model

Area of Science:

• Computer Science
• Electrical Engineering
• Networking

Background:

• Traditional medium access control (MAC) protocols struggle with the increasing scale and dynamic topology of wireless ad hoc networks (WANETs).
• Key challenges in modern WANETs include resource allocation, interference management, limited bandwidth, and energy constraints.
• Emerging WANET architectures require more intelligent solutions for efficient communication.

Purpose of the Study:

• To survey the literature on reinforcement learning (RL)-based MAC protocols for WANETs.
• To analyze existing state-of-the-art MAC protocols and their RL-based solutions within the MAC reference model.
• To identify future research directions for RL in high-performance MAC protocols.

Main Methods:

• Literature survey of RL-based MAC protocols for WANETs.
• Analysis of existing protocols based on the MAC reference model.
• Discussion of RL schemes and their application to MAC protocol components.

Main Results:

• Identified limitations of traditional MAC protocols in dynamic WANETs.
• Cataloged and analyzed current RL-based MAC protocols and their approaches.
• Highlighted challenges and potential solutions for RL integration in MAC protocols.

Conclusions:

• Reinforcement learning presents a promising approach to overcome limitations of traditional MAC protocols in WANETs.
• Further research into RL-enabled MAC protocols is crucial for achieving high performance in future wireless networks.
• Intelligent, adaptive MAC protocols are essential for efficient communication in complex, dynamic network environments.