Trust management in ubiquitous computing: A Bayesian approach

Designing a trust management scheme that can effectively evaluate the relationships among devices in pervasive computing environments is a challenging task. This paper continues the investigation of our recently proposed probabilistic trust management scheme for pervasive computing environments. We argue that in addition to allowing a device to find other appropriate devices with which to interact, while detecting those that are malicious, our trust management scheme is also capable of (1) allowing a device to judge the trustworthiness of another device it interacts with, while making a better use of the received recommendations and (2) behaving as expected when a device has little or enough experience of interactions with other devices and changes dynamically occur in the proportion of malicious devices. Simulation experiments are provided to assess the achievement of the stated goals, using some representative performance metrics.