Improved precision in a heuristic for particle-based and stochastic dynamic traffic assignment

Abstract

This work continues Flétteréd (2022) and refines an operational and efficient heuristic for the stochastic and particle-based dynamic traffic assignment problem. Experimental results indicate that the method performs advantageous when compared to alternative techniques that are applicable in the considered “agent-based” setting (variations of the method of successive averages (MSA) and the sorting technique of Sbayti et al. (2007)). Conveniently, the method requires no tuning parameters, not even an MSAlike step size rule.