Self-regulation Management in IoT Infrastructure Using Machine Learning

Abstract

AbstractThe emergence of the IoT concept introduced new opportunities and challenges in creating and functioning digital solutions that are IoT based. The IoT concept nature is a heterogeneous and distributed infrastructure that leads to some of the critical issues related to optimally manage several data streams and commands and avoiding high computation loading of nodes in the IoT infrastructure. Furthermore, the IoT infrastructure can dynamically change its state, location, and a number of IoT nodes caused by internal processes or environments. Moreover, the initial IoT concept was considered as a self-regulated system that each node can be connected with others directly. However, mostly there are highly prevalent architectures with a constant centralized computing node. In this regard, IoT infrastructure is required in mechanisms of optimal load balancing and self-intellectual management of IoT nodes. The paper proposes such mechanisms based on a machine learning approach that enables IoT infrastructure with rapid response to internal processes and the environment in real-time in uncertain conditions.KeywordsInternet of thingsMachine learningReinforcement learningIntelligent systemsDistributed systems