An Efficient Heuristic for AI-based Urban Traffic Control

Abstract

There is a growing interest in the use of Artificial Intelligence (AI) techniques for urban traffic control, with a particular focus on traffic signal optimisation. AI-based approaches demonstrated to be capable of dealing in real-time with unexpected or unusual traffic conditions, as well as with the usual traffic patterns. To effectively perform their task, AI approaches require the design of ad-hoc techniques, usually under the form of heuristics to guide the search process. In this paper, leveraging on the successful application of automated planning to urban traffic control, we introduce an innovative heuristic and test it using real-world historical data. The experimental analysis shows that the proposed heuristic allows to quickly generate high-quality traffic signal strategies, that outperform those generated by the SCOOT framework currently deployed in the considered region.