Abstract:

Trust is a critical concept in many networked systems including communication and sensor networks, smart power grids, smart transportation networked, human-robotic teams, the internet, economic networks, social networks and social networks over the internet. We develop a trust taxonomy and associated metrics related to direct and indirect trust. Indirect trust is directly related to reputations and recommendation systems. Direct trust is developed between agents or nodes that have previous direct transactions or collaborations. We develop a partially ordered semiring methodology for dynamic trust and distrust evaluation and evolution. We describe our results of applying this methodology in: trust aware collaboration, the impact of trust dynamics on security, team cohesion and opinion dynamics, network formation, social network clustering, healthcare informatics and economics, and networked collaborative control. We close with a brief description of future research directions.