Multi-objective optimization of freeway traffic flow via a fuzzy reinforcement learning method

A fuzzy approach to reinforcement learning in multi-objective optimization problem of freeway traffic flow control and dynamic route guidance is presented. The problem domain, a freeway network integration management application considers the efficiency and equity of system, is formulated as a distributed fuzzy reinforcement learning problem. The Gini coefficient is adopted in this study as an indicator of equity. The DFRL approach was implemented via a multi-agent control architecture where the decision agent was assigned to each of the on-ramp or VMS. The reward of each agent is simultaneously updating a single shared policy. The control strategy´s effect is demonstrated through its application to the simple freeway network. Analyses of simulation results using this approach show the equity of the system have a significant improvement over traditional control, especially for the case of large traffic demand. Using the DFRL approach, Compared with traditional methods, the network´s Gini coefficient has fallen by 30% or more.