Secure communication using θ-non-dominated sorting genetic algorithm.

With the growth in ad hoc devices, mobile ad hoc networks (MANETs) are extensively employed to realize different kinds of real-time applications. However, ad hoc devices suffer from numerous kinds of security attacks such as blackhole, grayhole, wormhole, etc. Therefore, an efficient trust management protocol is preferable. In this paper, a novel gradient boost classifier is proposed. Additionally, the hyper-parameters of gradient boost classifier are tuned using θ -non-dominated sorting genetic algorithm-III (θ NSGA-III). Optimized link state routing (OLSR) protocol is employed for experimental analysis. A multi-class dataset is collected by implementing the blackhole, grayhole, and wormhole attacks on the OLSR. Comparative analysis reveals that the proposed θ NSGA-III-based gradient boost classifier outperforms the competitive attack classification models. [ABSTRACT FROM AUTHOR]