A Q-Learning Clustering Approach for Load Balancing in Edge IoT Networks
Main Article Content
Abstract
The Internet of Things (IoT) has become one of the most crucial topics in technology, used in large fields such as healthcare systems, industrial automation, traffic monitoring, and smart city infrastructure. In IoT networks, smart devices are equipped with sensors and/or actuators that allow them to perceive and react to the environment, and this gives them the ability to collect and analyze massive volumes of data from various sources; the main challenges are managing and analyzing this data generated by the IoT devices. Clustering is a technique used in IoT to group devices together based on their characteristics, such as location, type, and usage patterns. It is an important tool for managing and making sense of the large amounts of data generate by IoT devices. It can help improve IoT networks' efficiency, security, and scalability. Several clustering protocols have been developed to enhance data collection performance in IoT networks. Most focus on partitioning networks with static topologies, which are not optimal in IoT networks. This paper proposes a dynamic clustering approach in IoT networks using reinforcement learning. Our solution can improve the load balancing. Our preliminary simulation results show that the proposed Q-learning solution can achieve higher load balancing scores compared to current static solutions.