Node Behavior Based Selfish Node Prediction System (NB-SNPS) In Manet Environment
Mobile Adhoc Network environment is a set of nodes spread over in the environment with no proper infrastructure. The presence of selfish nodes would be high in MANET environment which can drop the inbound packet without forwarding to the destination. In the earlier research technique, Multiple Constraint aware Glow worm Swarm Optimization Technique (MC-GSO) was introduced that intends to identify and prevent the selfish nodes behavior by saving the available resources. But in the existing and previous research works, selfish nodes are detected on the basis of node behaviors with regard to packet forwarding and behavior of receiving that would result in inaccurate decision making. In the presence of selfish node in the environment, it is necessary to carry out rerouting immediately for properly recovering the data items. These challenges get focused in the new research technique by presenting a new technique which is the Node Behavior based Selfish Node Prediction System (NB-SNPS). Here, at first, clustering is performed depending on the transmission range and the optimal cluster head selection is carried out by employing Intelligence Water Drops (IWD) algorithm. Afterwards, dynamic routing and rerouting is performed for immediate response by bringing in the Adaptive Routing method that tries to get the alternative routing path by which optimal data transmission can be guaranteed. Then, the packet forwarding behavior of each node is learned by making use of Reinforcement learning technique that can learn the node forwarding behavior. Depending on this information learned, packet routing is carried out in a dynamic manner, such that guaranteed packet delivery ratio with minimized delay can be attained. The performance evaluation of the proposed research is carried out in the matlab13 simulation environment, which proves that the new research methodology NB-SNPS can yield an optimal prediction rate of selfish nodes compared to the other available research techniques.