BVG Choice in Axelrod’s Tribute Model

Abstract

We consider Robert Axelrod’s tribute model at the light shed by the development of a choice framework called BVG (beliefs, values and goals). BVG agents use multi-varied, situated and individual rationality to perform adaptive choices in social environments. We then take the original experiment and carry it over using this choice scheme. By explicitly representing the agents preferences and the related decisional mechanisms, we are able to easily extend the experiment to tackle issues only implicitly addressed. The outcome of these new experiments is somewhat surprising, and seems to undermine the original conclusions. The experiments were carried out in a similar way to that of the original ones. However, we sustain that a more holistic attitude towards experimentation may be preferable. We conclude by arguing that the model of multiple values may help to principle the relation between experimenter and experiment, through the strengthening of bridges between the design of the experiment and the design of the agents involved in the simulation.