Microscopic modeling of the relaxation phenomenon using a macroscopic lane-changing model

Abstract

A crucial challenge faced by current microscopic traffic flow models is capturing the relaxation phenomena commonly observed near congested on-ramps: vehicles are willing to accept very short spacings as they enter the freeway, but “relax” to more comfortable values shortly thereafter. This paper introduces a framework to solve this problem using a macroscopic theory of vehicle lane-changing inside microscopic models. In this theory, lane changes take place according to a stochastic process that has been validated in the field, and whose mean value is a function of lane-specific macroscopic quantities. As a consequence, the lane-changing logic becomes very simple compared to existing microscopic lane-changing models, and requires only one extra parameter. The resulting microscopic model is validated with empirical data.