A multi-agent system for modelling urban transport infrastructure using intelligent traffic forecasts

Abstract

This paper describes an integrated approach for modeling transport infrastructure and optimising transport in urban areas. It combines the benefits of a multi-agent system, real time traffic information, and traffic forecasts to reduce carbon-dioxide emissions and offer flexible intermodal commuting solutions. In this distributed approach, segments of different modes of transport (e.g. roads, bus/tram routes, bicycle routes, pedestrian paths) are simulated by intelligent transport agents to create a rich multi-layer transport network. Moreover, a user agent enables direct interaction between commuters' mobile devices and the multi-agent system to submit journey requests. The approach capitalises on real-time traffic updates and historical travel patterns, such as CO2 emissions, vehicles' average speed, and traffic flow, detected from various traffic data sources, and future forecasts of commuting behaviour delivered via a traffic radar to calculate intermodal route solutions whilst considering commuter preferences. © 2013 Springer-Verlag.