Multi-agent model of trust in a human game

Individual-level trust is formalized within the context of a multi-agent system that models human behaviour with respect to trust in the Trust and Tracing game. This is a trade game on commodity supply chains and networks, designed as a research tool and to be played by human players. The model of trust is characterised by its learning ability, its probabilistic nature, and how experience influences trust. The validity of the trust model is tested by comparing simulation results with aggregated results of human players. More specifically the simulations show the same effects as human plays on selected parameters like confidence, tracing cost, and the trust update coefficient on observable game statistics like number of cheats, traces, certificates, and guarantees. © 2006 Springer-Verlag Berlin Heidelberg.