An Improved Uncertainty Reduction Scheme Based on Bayesian Prediction in MANETs

Formulating and evaluating trust is important for ensuring security and collaboration among the nodes in MANETs. The dynamic nature of mobile ad hoc networks may contribute to uncertainty in trust opinions. Uncertainty in trust opinions reflects the sufficiency of trust information obtained by a trustor node so that it can accurately compute the trust values of its neighboring nodes. Uncertainty can therefore be reduced by the collection and dissemination of more trust information proactively, exploiting mobility. But the infinite collection and dissemination process leads to communication and cost overhead. And when the trust convergence time increases due to the network size, the possibility of stale opinions also arises. To overcome these overhead, we propose to include the probabilistic Bayesian prediction of trust values, along with gathering of trust information at periodic intervals, before needed, thereby reducing frequent information collection and dissemination. This reduces the communication and cost overhead considerably. The prediction process also prevents aging of opinions when done at desired time intervals. Simulation results are presented to support the performance of the Mobility and Prediction Assisted Uncertainty Reduction Scheme (MPAURS).