Economic and Computational Aspects of Smart Transportation Systems

Traffic congestion in the main roads and junctions of our cities is a world-wide problem incurring huge financial losses to the global economy as well as other negative environmental and health effects. We propose to study an algorithmic and economic / game theoretic design of a smart traffic management system to dynamically schedule traffic based on real-time traffic conditions and value of time reports from drivers, aiming to minimize the economic damage of traffic congestion.

Many cities in the UK have various types of static congestion charges where ‘static’ refers to the fee being charge, to the hourly/daily limitations, and to the zone itself. Economically, this could be highly inefficient as various parts of the city could have different congestion patterns that a static fee for example could only partially resolve. A too-low fee will be ineffective while a too-high fee will significantly damage economic and social activities unnecessarily.

We propose to study the possibility of designing a system to dynamically set congestion fees and route traffic based on multiple factors. The proposed system will be composed of three parts: (1) an information module to disseminate knowledge about traffic across the network, enabling each road and junction to obtain a probabilistic forecast about the future arrival of cars; (2) a traffic light scheduling algorithm that, given this information and the information on the value of time of drivers arriving to junction, will decide on the green/red light schedule in order to minimize the aggregate cost of waiting in the junction; and (3) a payment scheme for charging payments from drivers, ensuring that drivers will not exaggerate (nor underestimate) their value of time reports. These payments will effectively create a dynamic toll system that will more efficiently ensure that drivers use roads in an economically efficient way, replacing today’s static toll systems. 

Such a dynamic system is advantageous since the resulting tolls depend on the actual route being taken, and drivers’ payments increase when using more congested junctions. The proposed research is to design the algorithmic and economic theory behind such a system, and to evaluate the obtained solutions via computer simulations. The proposed research will rely on a variety of techniques: methods of planning and scheduling from artificial intelligence, algorithmic methods from theoretical computer science, and methods for the design of incentive-compatible (“truthful”) mechanisms from game theory. Such an interdisciplinary interaction will benefit all these disciplines, as it will introduce new questions, new answers, and new tools and techniques, to all of them.

This project is offered as part of the Centre for Doctoral Training in Advanced Automotive Propulsion Systems (AAPS CDT).  The Centre is inspiring and working with the next generation of leaders to pioneer and shape the transition to clean, sustainable, affordable mobility for all. 

Prospective students for this project will be applying for the CDT programme which integrates a one-year MRes with a three to four-year PhD  

AAPS is a remarkable hybrid think-and-do tank where disciplines connect and collide to explore new ways of moving people. The MRes year is conducted as an interdisciplinary cohort with a focus on systems thinking, team-working and research skills. On successful completion of the MRes, you will progress to the PhD phase where you will establish detailed knowledge in your chosen area of research alongside colleagues working across a broad spectrum of challenges facing the Industry.  

The AAPS community is both stretching and supportive, encouraging our students to explore their research in a challenging but highly collaborative way. You will be able to work with peers from a diverse background, academics with real world experience and a broad spectrum of industry partners. 

Throughout your time with AAPS you will benefit from our training activities such mentoring future cohorts and participation in centre activities such as masterclasses, research seminars, think tanks and guest lectures. 

All new students joining the CDT will be assigned student mentor and a minimum of 2 academic supervisors at the point of starting their PhD. 

Funding is available for four-years (full time equivalent) for Home students. 

See our website to apply and find more details about our unique training programme (aaps-cdt.ac.uk)