Towards (Dis)Trust-Based Simulations of Agent Networks

$%675$&7 While the importance of trust in agent networks has been recognized for some time, the role of distrust has largely been ignored. In this paper, we argue that an explicit consideration of distrust and its complex interaction with individual trust and confidence in the network as a whole is necessary for the design and analysis of hybrid networks of human and machine agents. We propose a trust-confidencedistrust (TCD) model of agent network dynamics, and show how the dynamics of such networks can be modeled by extending Yu’s i* framework and combining it with the plan language ConGolog. This results in a simulation environment in which network processes and their dependence on trust, confidence in the networks and distrust can be studied.