The Challenge of Controlling the Queen Elizabeth Way: A Long Heavily Congested Urban Freeway with Dense Consecutive On-Ramp Bottlenecks

Benefits of ramp metering (RM) and variable speed limit (VSL) for a single or a few congested freeway ramp bottlenecks are well established in the literature; however, for a long heavily congested urban freeway with dense consecutive on-ramp bottlenecks, the application of RM and VSL is highly complex. Owing to short distances between consecutive bottlenecks, control actions at some ramps can interfere with other ramps. The main objective of this paper is to analyze and study the effectiveness of RM and VSL application on a full congested urban freeway as opposed to their application on one isolated bottleneck. In this paper, we apply both RM and VSL controllers on single-ramp and multi-ramp networks to compare and analyze the performance and effectiveness of each controller in each setting. We find that in the case of single-ramp bottleneck, both RM and VSL perform well under moderate traffic demand, which is expected and consistent with the literature. However, in heavily congested scenarios, the performance of VSL deteriorates while RM can preserve its performance. On larger multi-ramp networks, applying RM to resolve congestion at upstream bottlenecks causes more congestion at downstream bottlenecks. Further, if only VSL is applied, the lack of adequate distances between the congested bottlenecks and the high unmetered flows merging to the freeways from the on-ramps significantly reduce the efficiency and benefits of applying VSL. We demonstrate and quantify those challenges on a case study of the Queen Elizabeth Way (QEW) freeway in the Greater Toronto Area, Canada.