神戸大学六甲台第2キャンパス 自然科学総合研究棟3号館1階125室

Making route choice and traffic flow models more realistic

– Date and time: 7th November, 4:50pm – 6:20pm
– Speaker: Dr. Adam J Pel, TU Delft, the Netherlands
– Title: “Making route choice and traffic flow models more realistic, but not more complex”
– Venue: Room 125, Science & Technology Research Building No. 3, 1F
– Place : Rokkodai 2nd Campus, Department of Engineering, Kobe University.

In this seminar I will talk about route choice models and traffic flow models as these are used in road network modelling. Road network models simulate drivers’ behaviour and how their decisions are both based on, and collectively lead to, the emerging traffic flows and congestion conditions.
I will present several existing types of models, from basic to complex, and discuss their underlying assumptions on traffic behaviour and their suitability for various modelling applications. I will present in more detail a few recent studies at Delft University of Technology on: a route choice model that incorporates dynamic rerouting behaviour; a static traffic flow model that incorporates ‘dynamic’ traffic flow congestion; and a first-order traffic flow model that incorporates some second-order traffic flow phenomena.Bio:
Dr. Adam Pel is Assistant Professor at Delft University of Technology in the Netherlands. His main research field is the resilience of road transport systems. Pel’s research group studies, on the one hand how stochastics, uncertainty, dynamics and disruptions affect transport systems, including emergencies and evacuations, and on the other hands how network design, contingency planning, and mobility and traffic management can be strategically used to increase resiliency. In his research, he often uses network modelling as research method. These models are used to assess the dynamic performance of road transport systems regarding: human factors, infrastructure, services, technologies, policies, control measures and information flows. Pel’s research group develops models for more behavioural realism, faster computation, better use of (new) data, higher precision and accuracy. Furthermore, Pel works part-time at Fileradar, a university-spinoff company, where he is lead engineer for Fileradar’s predictive data analytics, used for traffic monitoring, information and control.