A Novel Machine Learning Algorithm for MANET Attack: Black Hole and Gray Hole

Abstract

Mobile ad hoc networks (MANETs) are a class of wireless networks that can be operated without a fixed infrastructure. Due to the dynamics of decentralised systems, these networks are prone to different attacks like Black Hole Attack (BHA) and Gray Hole Attack (GHA). The basic requirement in this network is that all nodes are trusted nodes, but in a real-life scenario, some nodes may be malicious, so instead of transferring the data packet to the destination, it drops the data packet. Organisations have some ideas for preventing this attack but can fail due to improper methods, so the attack must be identified and addressed. This article uses the deep learning algorithm concept with a mutation-based artificial neural network (MBNN). It uses a swarm-based Cluster-Based Artificial Bee Colony (CBABC) optimisation technique to protect this network from BHA and GHA attacks. The proposed models performance has been improved by selecting the appropriate and best node for sending data packets. We have demonstrated experimental results suggesting that the proposed protocol outperforms existing work in the case of black and gray hole attacks.