Routing in Mobile ad Hoc Networks Using Machine Learning Techniques

Abstract

The significance of Mobile Ad Hoc Networks (MANETs) has increased in the current period due to the proliferation of mobile devices and the emergence of Internet of Things (IoT) applications. The function of routing in decentralized wireless networks is of paramount importance in facilitating efficient data transfer. Conventional routing techniques need help in effectively adjusting to the dynamic changes in network topology and the constraints imposed by restricted network resources. To tackle these concerns, a Machine Learning-Based Optimized Routing Algorithm (ML-ORA) is proposed in this research. ML-ORA has been developed to offer adaptable and intelligent routing solutions, especially in MANET. This is achieved via the integration of parameter settings, the use of a Hybrid Particle Swarm Optimization (HPSO) algorithm for Cluster Head (CH) selection, the inclusion of a clustering stage, and the incorporation of a k-Nearest Neighbors (k-NN)-based intrusion detection system. The performance of ML-ORA is assessed by conducting simulations using Network Simulator 3 (NS-3), demonstrating its efficacy in dynamic network settings. The findings exhibit a latency of 15 milliseconds, a packet delivery ratio of 95%, a network throughput of 4Mbps, and an energy efficiency of 85%. ML-ORA presents a potentially advantageous resolution for tackling the obstacles encountered in MANET routing, thereby creating opportunities for enhanced efficacy and security in the transmission of data inside dynamic wireless networks.