Construction of an Intrusion Relieved Communication Channel using a Hybrid Optimized Algorithm

In wireless network, several nodes try to communicate with each other by interchanging information between them. These wireless networks are highly subject to intrusion from unauthorized attackers. There are many intrusion detection systems to find the intrusion in these kinds of networks which have been proposed and implemented successfully. Yet there has been minimum amount of interest in developing an architecture which is relieved from intrusions, instead of just detecting them. In the previous works, we have constructed architecture for an intrusion relieved wireless network based on trust level of every node in the network and it has performed well too. The performance level of this intrusion relieved architecture could be boosted by using a hybrid technique to detect intrusion. In this paper, a hybrid technique comprising modified RotBoost algorithm, AdaBoost algorithm and Artificial Bee Colony (ABC) optimization algorithm to estimate the trust level of every node in upcoming instants is proposed. In this proposed hybrid technique, the ABC algorithm and the modified RotBoost learning algorithm [22] are executed in parallel manner. The modified RotBoost learning algorithm is used to find the fitness value of the predefined node which is used to compare with the nodes fitness from the employee bee phase of ABC algorithm. This comparison process takes place in the onlooker bee phase of ABC algorithm. At last in the scout bee phase in ABC is used to find the trusted node to check the intrusion. As per the information obtained from the scout bee phase the intrusion detector detects the intrusion free path. ABC is a naturally inspired optimization algorithm based on the behavior of honey-bee swarms and it is easy to implement. This hybrid technique comprising AdaBoost in ABC could detect the intruding nodes more effectively without making any adverse change to computation time of the existing architecture.