基于机器学习的移动自组网MAC协议研究综述

A Survey on MAC Protocol of Mobile Ad Hoc Network Based on Machine Learning

MAC协议设计是移动自组网的一项重要关键技术,主要通过有效的多节点协商机制,实现各节点对空间、时间、频率等有限资源的合理共享,完成各节点的随机接入和资源分配功能。随机接入和资源分配问题通常可用最优化和马尔科夫决策问题表述,机器学习可成为提升移动自组网MAC协议性能的有力手段之一。该综述在分析机器学习和移动自组网MAC协议特点基础上,结合现有研究成果,介绍了基于机器学习的关键技术原理和仿真性能,展望了未来研究方向。现有研究成果表明,基于机器学习的MAC协议设计性能改善显著。

MAC protocol is a key technology for mobile ad hoc networks.It mainly implements reasonable sharing of limited resources such as space,time,and frequency for various nodes through an effective multi-node negotiation mechanism,and achieves random access and resource allocation of each node.Random access and resource allocation problems can usually be modeled by optimization and Markov decision problems.Machine learning can become one of the powerful means to improve the performance of MAC protocols in mobile ad hoc networks.Based on the analysis of the characteristics of machine learning and mobile ad hoc network MAC protocols,this paper introduces the key technical principles and simulation performance of machine learning based on existing research results,and looks forward to future research directions.Existing research show that the performance of MAC protocols based on machine learning has improved significantly.