Modeling Social Deviance in Artificial Agent Societies

Abstract

Rule-governed artificial agent societies consisting of autonomous members are susceptible to rule violations, which can be seen as the acts of agents exercising their autonomy. As a consequence, modeling and allowing deviance is relevant, in particular, when artificial agent societies are used as the basis for agent-based social simulation. This work proposes a belief framework for modeling social deviance in artificial agent societies by taking into account both endogenous and exogenous factors contributing to rule compliance. The objective of the belief framework is to support the simulation of social environments where agents are susceptible to adopt rule-breaking behaviors. In this work, endogenous, exogenous and hybrid decision models supported by the event calculus formalism were implemented in an agent-based simulation model. Finally, a series of simulations was conducted in order to perform a sensitivity analysis of the agent-based simulation model.