Cooperative learning hybrid agents for traffic management and control

Abstract

This paper presents the application of fuzzy-neuro-evolutionary hybrid system with online reinforcement learning for intelligent road traffic management and control. The hybrid system presents different methodologies in knowledge acquisition, decision-making, learning and goal formulation with the use of a three-layered hierarchical, distributed agent architecture. Distributed and hierarchical fuzzy knowledge acquisition allows different levels of perception to be derived for the same traffic situation by the intelligent agents. Agents' perceptions can be changed with the use of online reinforcement learning. Initial experimental results show that the implementation of the hybrid agents in a signalized network generally yields better traffic performance when compared to a network without the agents. The probability of a traffic network evolving into pathological states with oversaturation is also reduced with the implementation of the agents.