QoS Sensitive Routing Algorithm with Link Quality Prediction in SDN-based Ad Hoc Networks

The mobility of the nodes may cause route failures frequently in the SDN-based ad hoc networks. Routes interruption causes the Quality of Service (QoS) of the users can't be satisfied, as well as leading the performance of the wireless mobile network to degrade. In this paper, a QoS sensitive routing algorithm with link quality prediction in SDN-based ad hoc network (QSRLQP) is proposed. The algorithm includes two parts, which is link quality prediction based on XGBoost (LQPXGB) and routes generation based on minimum cost tree (RGMCT). LQPXGB uses the interior point method to optimize the leaf node score of each decision tree to further predict the link quality. RGMCT uses the minimum cost tree method to get the best route, which can achieve the highest packet delivery ratio (PDR) with guaranteed delay. The simulation results show that, compared with existing routing algorithm, QSRLQP achieves a significant performance improvement in terms of PDR, throughput and end-to-end delay.