2021/05/20 by Hossam Farag, Farag, Hossam, Čedomir Stefanović +1
Computer Science · Engineering · #Advanced MIMO Systems Optimization #Energy Harvesting in Wireless Networks #FOS: Computer and information sciences #Networking and Internet Architecture (cs.NI) #Software-Defined Networks and 5G
paper · pdf · doi:10.48550/arxiv.2105.09678
openalex publication_date 2021/05/20 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
The innovative services empowered by the Internet of Things (IoT) require a\nseamless and reliable wireless infrastructure that enables communications\nwithin heterogeneous and dynamic low-power and lossy networks (LLNs). The\nRouting Protocol for LLNs (RPL) was designed to meet the communication\nrequirements of a wide range of IoT application domains. However, a load\nbalancing problem exists in RPL under heavy traffic-load scenarios, degrading\nthe network performance in terms of delay and packet delivery. In this paper,\nwe tackle the problem of load-balancing in RPL networks using a\nreinforcement-learning framework. The proposed method adopts Q-learning at each\nnode to learn an optimal parent selection policy based on the dynamic network\nconditions. Each node maintains the routing information of its neighbours as\nQ-values that represent a composite routing cost as a function of the\ncongestion level, the link-quality and the hop-distance. The Q-values are\nupdated continuously exploiting the existing RPL signalling mechanism. The\nperformance of the proposed approach is evaluated through extensive simulations\nand compared with the existing work to demonstrate its effectiveness. The\nresults show that the proposed method substantially improves network\nperformance in terms of packet delivery and average delay with a marginal\nincrease in the signalling frequency.\n