Author(s)

G. E. Cantarella, P. Velonà & A. Vitetta

Abstract

In this paper a model and a procedure for signal setting design with demand assignment are reported. The model is developed within a day-to-day dynamic framework where demand assignment is dealt with through deterministic (or stochastic) process models. The system of models, based on a what-to approach, generates signal timings taking into account users path choice behaviour as well as stability constraints. The main contribution in this paper is the specification of the heuristic procedure for signal setting optimization, based on genetic algorithm. The procedure is tested on a real scale test site with the objectives to validate the system of models and the optimization procedure. A sensitivity analysis with respect to demand level is also briefly discussed. Keywords: signal setting, traffic control, day-to-day dynamic. 1 Introduction This paper deals with the design of signal setting in urban areas. This problem can not be solved without considering the user behaviour at least at route choice level. The resulting problem belongs to the class of problem of transportation supply design with assignment [1–4]. Two main approaches have been proposed with respect to the method for demand assignment (with probabilistic path choice behaviour): equilibrium assignment, effectively approached by fixed-point models; day-to-day dynamic process models, derived from time-discrete non-linear dynamic system theory.

Keywords

signal setting, traffic control, day-to-day dynamic.