High-speed adaptive and efficient ad hoc on-demand distance vector for secure routing in mobile ad hoc networks using reinforcement learning and artificial bee colony optimization

Abstract

Mobile ad hoc networks (MANETs) have provided the opportunity to establish a temporary mobile network with no fixed centre or support. However, it is a fact that the security of routing systems is greatly affected by the inherent features of MANETs. Ad hoc on-demand distance vector (AODV) is susceptible to attacks such as black hole, grey hole, and wormhole attacks, despite the fact that it is the most prevalent routing protocol in MANETs. This study tries to offer solutions to these issues by enhancing the security and efficiency of the routing process in MANETs through the formulation of the Reinforcement Learning (RL)–Artificial Bee Colony (ABC)–AODV protocol, which is a combination of RL and ABC Algorithm in AODV protocol. According to the network's behaviour, the ABC algorithm selects the optimal path from a list of choices based on a number of factors. The suggested RL–ABC–AODV technique also contains dynamic and adaptive aspects in which the Q-learning and the network's current state are used to adjust parameters in real-time. The RL–ABC–AODV protocol achieved a packet delivery ratio of 99.45%, an end-to-end delay of 0.02 s, a throughput of 250.6 Mbps, and a routing overhead of 0.8, significantly outperforming existing methods.