An Evolutionary Agent-Based Approach to Social Norm Emergence in Traffic Signals

Abstract

For multi-agent systems such as our society, formation and maintenance of some “norms” are important to keep the system efficient and stable. However, mechanisms that such norm are formed, maintained and collapsed are not clear and very fruitful but difficult open problem. Agent-based simulation is a promising approach to this problem. This paper presents a compact framework of simulation with learning agents in traffic signal system as an example of norm formation. Through the implementation of the framework with Genetic Algorithm and some preliminary experiments, the potential of the framework is shown.