Optimal design of intersections in an urban transport network
Drs. Ing. E. Bezembinder
The project considers an urban transport network with controlled and non-controlled intersections, and is about the optimal design of all intersections in the network, where transport demand is given. The research approach is a so-called multi-objective network design problem, where decision variables are related to the design of all intersections. In case of a controlled intersection the layout of the intersection and the configuration of the control plan are decision variables, in case of an uncontrolled intersection the type of intersection (e.g. roundabout or regular intersection) with the layout are decision variables.
Objective functions are related to network efficiency, the emissions of GHG, traffic safety and noise. The solution approach is bilevel, where the upper level deals with the network design and the lower level with individual travellers that try to optimize their own situation. For the lower level a modified version of the traditional Frank-Wolfe equilibrium approach is developed. In the upper level a modified form of a genetic algorithm is developed where so-called design grammers are included.
Bezembinder, E.M., Wismans, L.J.J. & Berkum, E.C. van (2015). Using decision trees to determine junction design rules. In TRB (Ed.), --- Proceedings of the 94th Annual Meeting of the Transportation Board, 11-15 January 2015, Washington DC, USA. (on CD ROM). --- (pp. 1-16). Washington DC: TRB.
Bezembinder, E.M., Wismans, L.J.J. & Berkum, E.C. van (2014). Using decision trees to determine junction design rules. In --- Proceedings of the European Transport Conference, September 29th - October 1st, 2014, Frankfurt, Germany. --- (pp. 1-12). Frankfurt: ETC.