A trust evaluation model using controlled Markov process for MANET

A distributed trust evaluation model is presented for MANETs by which uncertainties of trust are transformed into probability vectors giving the probability distribution of trust levels. The system evolves over time as a finite-state Markov process with variant transition matrixes. We attempt to predict the trustworthiness values of entities that are determined by their inherent error patterns. The Markov process is associated to a Bonus-Malus System and controlled by the estimated error patterns of individual entities involved. As well, an iteration algorithm is designed to prevent inaccurate predictions for trust values because of the properties of Markov process. The simulation results demonstrate that our model is able to predict local trust successfully for entities in MANETs by estimating their actual error patterns accurately.