How to do social simulation in logic: modelling the segregation game in a dynamic logic of assignments

Benoit Gaudou (1) , Andreas Herzig (2, 3) , Emiliano Lorini (2, 3) , Christophe Sibertin-Blanc (1)

1 IRIT-SMAC - Systèmes Multi-Agents Coopératifs
2 IRIT-LILaC - Logique, Interaction, Langue et Calcul
3 CNRS - Centre National de la Recherche Scientifique

The aim of this paper is to show how to do social simulation in logic. In order to meet this objective we present a dynamic logic with assignments, tests, sequential and nondeterministic composition, and bounded and non-bounded iteration. We show that our logic allows to represent and reason about a paradigmatic example of social simulation: Schelling’s segregation game. We also build a bridge between social simulation and planning. In particular, we show that the problem of checking whether a given property P (such as segregation) will emerge after n simulation moves is nothing but the planning problem with horizon n, which is widely studied in AI: the problem of verifying whether there exists a plan of length at most n ensuring that a given goal will be achieved.

Mots clés: segregation game; dynamic logic; complexity

Domaines: Intelligence artificielle [cs.AI]

Benoit Gaudou, Andreas Herzig, Emiliano Lorini, Christophe Sibertin-Blanc. How to do social simulation in logic: modelling the segregation game in a dynamic logic of assignments. International Workshop on Multi-Agent Systems and Agent-Based Simulation (MABS 2011), May 2011, Taipei, Taiwan. pp.59-73, ⟨10.1007/978-3-642-28400-7_5⟩. ⟨hal-03470293⟩