Agent-based experimental economics in signaling games

Abstract

In this paper we built up an agent-based model inspired by human-subject behaviour in a signaling game experiment. The behavioural patterns of players observed in the experiment are classified in terms of attitudes, emotions and frugal and fast heuristics. We build up an agent-based model where artificial agents are endowed with these cognitive patterns by means of an endorsement scheme. We show the relationship between cognition and Multi-Agent interaction. We validate the model since it reproduces well the behavioural patterns observed in the human-subjects experiment. © 2008 Springer-Verlag Berlin Heidelberg.