A Multi-Agent Model Inspired in Soft-Power

Abstract

Soft-power is a relatively new player in the foreign affairs sciences. Instead of using coercion, characteristic of hard-power, it grows on the attraction of shared values. Multi-agent simulations are a tool of choice to study the social nature of soft-power, where realistic scenarios requiring heterogenous agents, acting rationally with respect to their own, individual, rewards can be explored.

The work described here is an experiment where deliberation results from a simple individual markov decision process that tries to incorporate elements inspired by soft-power. The resulting policies are then analyzed and tested in a virtual multi-agent simulation.