应用于战术边缘网络的随机森林路由算法  

Routing Algorithm Based Random Forest in Tactical Edge Network

作　　者：殷丽丽 钱良[1] 吴冠钰 YIN Li-li;QIAN Liang;WU Guan-Yu(The School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《计算机仿真》2020年第6期293-297,共5页Computer Simulation

摘　　要：战术边缘网络几乎很难保证端到端的链路,只能利用节点的移动性和存储能力来实现通信。现有的路由协议中大多只关注某一特定的路由因素(如相遇概率)且中继节点只能被动接受消息,很难适应机动性强的战术边缘网络。提出了一种动态的基于随机森林算法的路由算法(routing algorithm based Random Forest,RAbRF),中继节点能根据系统网络的各种因素对所要转发的消息基于随机森林算法进行排序,从而选择最偏向的消息进行存储转发。RAbRF算法与基于概率的路由算法进行了仿真对比,结果表明,提出的算法能够有效改善消息传递成功率,传输时延。Tactical edge networks which are almost impossible to guarantee end-to-end links, can only use the mobility and storage capabilities of nodes to achieve communication. Most of the existing routing protocols only focus on a specific routing factor, such as the probability of encounter. What’s more, relay nodes can only passively accept messages. It is are difficult for them to adapt to the tactical edge network with strong mobility. A dynamic routing algorithm based on Random Forest(RAbRF) is proposed in the paper. The relay node can sort the messages to be forwarded based on the random forest algorithm according to various factors of the system network, thus select the most biased message which will be stored and forwarded. The RAbRF protocol was compared with the probability-based routing protocol, and the results show that the proposed protocol can effectively improve the message delivery success rate and transmission delay.

关 键 词：战术边缘网络 随机森林算法 消息副本数量 消息优先级 最短期望路径 

分 类 号：TJ01[兵器科学与技术—兵器发射理论与技术]