Agent-Based Modeling of Traffic Systems Using Modelica

A microsimulation model of traffic described using the Modelica language in combination with the Modelica libraries developed by the authors, ABMLib and CellularAutomataLib2, is presented. Modelica facilitates the description of equation-based models using the physical modeling paradigm, and its combination with the discrete-event and agent-based modeling functionality of the developed libraries constitutes a powerful and versatile tool. The functionality of ABMLib is extended to facilitate the description of individual agent behavior as a combination of equation-based and discrete-event models. In this way, vehicle dynamics and fuel consumption are modeled using equations, while driver behavior is described using an agent-based model. Vehicles move along the road, which is described as a combination of interconnected segments that constitute the environment for the agents. A cellular automaton is used to model the CO2 emissions of vehicles to the environment. A model that represents two consecutive road segments with a semaphore between them is used to illustrate the provided functionality.