High occupancy vehicle (HOV) lanes can affect the capacity of freeway bottlenecks through both an under-utilization effect and a disruption effect. An under-utilized HOV lane obviously reduces its own discharge rate at the bottleneck. But lane changes in and out of the HOV lane can also disrupt the flow on the adjacent general purpose (GP) lanes, and reduce their capacity. Bottleneck capacity reductions, arising from the combination of both effects, are undesirable because they increase vehicle-hours of travel. This paper shows with systematic simulations that the disruption effect is not significant at isolated bottlenecks. If anything, HOV lanes slightly improve GP flow at these locations. Reductions in GP-capacity were found only in highly idealised situations without bottlenecks. Thus, it appears that the total discharge rate of bottlenecks with and without HOV lanes can be conservatively analysed assuming that the capacity of the GP lanes is constant. Using this assumption, the paper then shows how to estimate total bottleneck capacity, and how to deploy HOV lanes without creating new bottlenecks or changing the total flow through existing ones. This milestone is sometimes sufficient to guarantee no change in the total vehicle-hours of travel and, hence, a reduction in people-hours. The paper also introduces a dynamic operating strategy for HOV lanes that significantly increases a bottleneck's total discharge rate. This suggests that dynamic strategies can be used to reduce not just people delay but also vehicle-hours and their externalities (Author/publisher)
Abstract