On the Automatic Synthesis of Social Laws for Mobile Robots. A Study in Artificial Social Systems

Abstract

We introduce an algorithm for the automatic synthesis of social laws for mobile robots. Our algorithm generates useful social laws for any 2/connected grid-like environment with arbitrary obstacles. We prove that the social laws generated by our algorithm enable the agents to achieve their goals while preventing collisions. Moreover, computer simulations show that the social laws generated by the algorithm lead to efficient behaviour in a large set of environments. Our work bridges the gap between the work on the automatic synthesis of social laws in abstract models, and work on hand-crafting social laws for a particular domain.