An ABC-AIS Hybrid Approach to Dynamic Anomaly Detection in AODV-Based MANETs

Network topology in mobile ad hoc networks (MANETs) changes rapidly due to the mobility of nodes. Hence, an important challenge for these networks is to develop an approach which can detect anomalies in network traffic with high accuracy despite the dynamic changing of network topology. In this paper, we present a hybrid approach based on the artificial bee colony (ABC) and negative selection (NS) algorithms, called BeeNS, for dynamic anomaly detection in AODV-based MANETs. In the approach, every node first extracts a set of feature vectors of its own normal network traffic. Each feature vector is represented by a hypersphere with fixed radius in the feature space. It then applies the NicheNABC algorithm to generate a set of mature negative detectors for covering of the nonself space. The negative detectors, represented by hyperspheres with variable radii, are used to detect anomalies in network traffic. It eventually updates the mature negative detectors by one of two methods of partial updating or total updating. In order to evaluate the performance of BeeNS, we simulated some routing attacks in AODV-based MANETs using the NS2 simulator. The experimental results demonstrate that BeeNS achieves comparable or better performance than the approaches previously reported in the literature, in terms of detection and false alarm rates.