Preventing bluff agent invasions in honest societies

Abstract

Frequently debated issues in the domain of game theory involve the issue of signalling strategies used in order to resolve conflicts between agents over indivisible resources and to reduce the costly outcomes associated with fighting. Signalling behaviour, used by agents of different strengths, to aid resource acquisition was modelled using an artificial life simulation environment. Honest signalling and the bluff strategy based on Enquist/Hurd's adapted pay-off matrix (1997) were evaluated relative to different proportions of resident strong agents capable of imposing a 'punishment' cost on bluffer agents.We found that in order for honest signalling to be immune to invasion by a bluff strategy, the number of punishment enforcers in the society must be high. Additionally, the number of punishment enforcers is more influential in preventing bluff agent invasions than the severity of punishment.