Transportation Policy Evaluation Using Minority Games and Agent-Based Simulation

摘要

Traffic congestion is an issue regarding the vitality of cities and the welfare of citizens. Transportation systems are using various technologies to allow users to adapt and make different decisions towards transportation modes. Modification and improvement of these systems affect the commuters’ perspective and social welfare. In this study, the effect of road flow equilibrium on commuters’ utilities with different types of transportation modes will be discussed. A simple network with two modes of transportation will be illustrated and three different cost policies were considered to test the efficiency of reinforcement learning in commuters’ daily trip decision-making regarding time and mode. The artificial society of agents is simulated to analyse the results.