In June, we organised our third webinar!

This webinar focused on transport network optimisation and arbitration models. It was moderated by Lakshya Pandit from Rupprecht Consult, a TANGENT partner, and had two speakers: Arka Ghosh from University of Deusto and Marios Giouroukelis from National Technical University of Athens.

‘Transport network optimisation problems and techniques’ – Dr. Arka Ghosh (DEUSTO)

Today, traffic policy and response plans have become challenging due to the multimodal nature of travel. These involve diverse stakeholders in the transport network management making the execution of decision-making complex. Public authorities must orchestrate both current and future transport modes to manage multimodal traffic and this reflects upon the need for new tools that support decision makers.

Approach using Computer Optimized Desgin (COD): TANGENT project uses COD to generate transport policies and/or interventions to support traffic management. All methodologies are based on AI and transport modelling consensus reaching mechanisms.

Two uses cases were presented:

1 – Integration of demand responsive transport and transit modes: In TANGENT, COD is applied to determine the optimal or near-optimal DRT fleet and operational routes to serve the first/last-mile of an area. Performance indicators include percentage of demand served, total emissions produced by DRT fleet, operational costs of the DRT service, and average travel time for the passengers.

2 – Synchronisation of public transport and traffic control: This use case is specific to urban scenarios where traffic lights and line-specific frequency/capacity of PT are optimised. The TANGENT approach for this use case includes consideration of certain planned or unplanned events, the frequency and/or capacity of some surface PT lines, and traffic control plan re-adjustment.

Methodology testing: In the Athens case study, the DRT-integration methodology was tested with 6 and 16 DRT vehicles. Similarly, the synchronisation of PT and traffic control was tested in a small network near Old Trafford stadium in Manchester.

‘Development of arbitration models’ – Marios Giouroukelis (NTUA)

Marios presented development of arbitration models focusing on the consensus mechanism.

Consensus mechanism: The consensus mechanism is coupled with optimisation problems within TANGENT. It involves gathering opinions through pairwise evaluations/preferences of different transport network objectives, such as service performance, network efficiency, economic performance, and environmental performance.

Stakeholder communication: Understanding how stakeholders communicate with one another is crucial. This includes asking stakeholders to nominate possible partners and provide the nature of interactions. A stakeholder graph is built for each problem using a data-driven method, the Bayesian network.

Decision support tool: A decision-support tool was created to facilitate decision-making through realistic modelling of usual steps in reaching decisions. This involves state-of-the-art, data-driven models trained on actual stakeholder preferences following a modular approach. The preference part of the preference-based optimisation approach is adopted in WP5 within the TANGENT project.

Resources

Presentations available here and recording here.

For further information, the Webinar Meeting minutes are available on the TANGENT Forum. If you’re interested to join the Forum, you can contact Lakshya Pandit.