Optimization of Variable Approach Lane Use at Isolated Signalized Intersections

Abstract

The variable approach lane (VAL) at an isolated signalized intersection is an effective way to handle the variation of arrival traffic at intersection approaches by reallocating the space resource according to the time-varying demand volumes. This study developed a model that simultaneously optimizes VAL use and signal timings, with the objective of minimizing the average intersection delay. The model constraints mainly include three parts: VAL use, signal timings, and degree of saturation. The application of the model was demonstrated through two numerical examples. The findings support the notion that the optimization algorithm can effectively reduce traffic delay. However, the benefits of VALs for intersection performance varied under different traffic conditions. One VAL installed on a single approach can effectively reduce cycle length and average delay when the left-turn traffic on that approach is greater than 600 passenger cars units per hour (pcu/h). VALs installed on two opposing approaches provided optimal solutions in 58 of 81 traffic combinations. The average delay decreased by more than 20%, with the exception that no VAL was needed when the left-turn traffic in both approaches was 200 pcu/h. Simulation of the scenarios in Vissim showed similar results of the model under different conditions, which verified the model’s validity. The results suggest that the availability of VALs provided the intersection with improved capacity to deal with large traffic fluctuations.