Abstract

This paper presents a trust-based evolutionary game model for managing Internet-of-Things (IoT) federations. The model adopts trust-based payoff to either reward or penalize things based on the behaviors they expose. The model also resorts to monitoring these behaviors to ensure that the share of untrustworthy things in a federation does not hinder the good functioning of trustworthy things in this federation. The trust scores are obtained using direct experience with things and feedback from other things and are integrated into game strategies. These strategies capture the dynamic nature of federations since the population of trustworthy versus untrustworthy things changes over time with the aim of retaining the trustworthy ones. To demonstrate the technical doability of the game strategies along with rewarding/penalizing things, a set of experiments were carried out and results were benchmarked as per the existing literature. The results show a better mitigation of attacks such as bad-mouthing and ballot-stuffing on trustworthy things.