Improving the Scalability of Automated Planning-based Vehicle Routing via Smart Routes Identification

Due to growing urbanisation, traffic infrastructures have to accommodate increasing demands of traffic volume. One promising way for supporting a better exploitation of traffic networks is vehicle routing, that can distribute traffic from congested links to under utilised ones. Automated Planning techniques, a research field of Artificial Intelligence, have demonstrated to be a suitable approach for performing effective centralised traffic distribution. However, a main weakness of this class of approaches is the limited scalability to large and complex networks. In this paper, we aim to improve the scalability of automated planning techniques for urban traffic distribution by introducing an approach for the identification of routes to be considered. The proposed technique can significantly improve the planning capabilities by simplifying the complexity of the urban network to be considered, as demonstrated by our extensive experimental analysis performed on realistic traffic data on part of the New York and Sydney urban areas.