Optimizing Data Routing in Wireless Sensor Networks: A Hybrid Approach Using Machine Learning and Tabu Search Mechanisms

This work presents a new technique for routing data in Wireless Sensor Networks (WSNs), where the extensive sensor nodes are capable of wireless communication, computing, and sensing, mainly for environmental data collection purposes. In WSN setup, the nodes, that transmit data to a central base station, face a vital challenges in data transmission, routing, and congestion management. Various prior protocols have been suggested for these tasks; however, they still have some limitations related to throughput, energy efficiency and network lifetime. Consequently, this paper focuses on optimizing energy use, improving network longevity, and enhancing Quality of Service (QoS) metrics such as active node count and data throughput. To this end, we integrate a machine learning (ML) mechanism with a traditional routing technique, where ML is employed to dynamically adjust routing decisions for better data transmission and congestion management, meanwhile, the Tabu Search algorithm identifies congested routes. Noting that a congestion alert is triggered when the traffic load on a route hits 70 percent, so, this activates our proposed approach to find alternative pathways to the Cluster Head (CH).