2020/12/31 by George Stamatakis, Stamatakis, George, Νικόλαος Παππάς +5
Computer Science · Engineering · #Age of Information Optimization #FOS: Computer and information sciences #IoT Networks and Protocols #IoT and Edge/Fog Computing #Machine Learning (cs.LG) #Networking and Internet Architecture (cs.NI)
paper · pdf · doi:10.48550/arxiv.2012.15548
openalex publication_date 2020/12/31 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Internet of Things (IoT) with its growing number of deployed devices and\napplications raises significant challenges for network maintenance procedures.\nIn this work, we formulate a problem of autonomous maintenance in IoT networks\nas a Partially Observable Markov Decision Process. Subsequently, we utilize\nDeep Reinforcement Learning algorithms (DRL) to train agents that decide if a\nmaintenance procedure is in order or not and, in the former case, the proper\ntype of maintenance needed. To avoid wasting the scarce resources of IoT\nnetworks we utilize the Age of Information (AoI) metric as a reward signal for\nthe training of the smart agents. AoI captures the freshness of the sensory\ndata which are transmitted by the IoT sensors as part of their normal service\nprovision. Numerical results indicate that AoI integrates enough information\nabout the past and present states of the system to be successfully used in the\ntraining of smart agents for the autonomous maintenance of the network.\n