Implementing Fire Fly Me-ta-Heuristic Algorithm For Identifying, Detecting, And Eliminating Coopera-tive Blackhole Attacks in MANET

Abstract

With the development of IoT and wireless sensor networks, the usage of MANET is increasing worldwide. Due to the non-linear and dynamic nature of the MANET, various attacks are made on the network. MANET networks are widely subjected to black hole attacks and denial of service attacks. It is easily detectable through various protocols. However, in the case of a larger network, the attack affects multiple nodes, leading to the cooperative black hole attack. Various researchers have designed various security mechanisms and protocols to tackle the cooperative black hole attack made on the MANET. The frequently attacked regions or the black hole regions need to be identified to tackle these attacks in the network. Various machine learning and deep learning algorithms are used to identify that region. Among the various machine learning algorithms used to detect these attacks, the classification algorithms help to identify the black hole regions. In this paper, we, as a collaborative team, propose a hybrid machine learning algorithm that uses an XGBoost classifier and a meta-heuristic approach to predict the attacks. The Firefly algorithm is used as the meta-heuristic approach to detect black holes. The proposed model is experimented with an open-source dataset, and the results are discussed in detail. The results show that the hybrid machine-learning model provides better accuracy in the cooperative black hole detection process.