Comprehensive Assessment of Managed Lane Performance and Characteristics
Managed lanes that include high occupancy vehicle (HOV) and high
occupancy and toll (HOT) lanes have been conducted for decades. Although
being regarded as efficient and sustainable transport, managed lanes
face such undiscovered issues as their performance regarding speed
dispersion, equilibrium relationships between managed lanes and general
purpose (GP) lanes in terms of speed and level of service, and joint
evaluation of managed lane elements like eligibility, access control,
and pricing. The goal of this dissertation is to provide theoretical and
practical approaches to assessing managed lane operations under four
modules, namely speed dispersion analysis, speed equilibrium analysis,
lane management hot spot analysis, and optimal managed lane policy
assessment. The first module correlates speed dispersion with the
fundamental traffic flow parameters, and reveals that the coefficients
of variation of speed for HOV and GP lanes are exponential with
occupancy, negative exponential with space mean speed, and two-phase
linear to flow, while the standard deviations of speed for both lanes do
not display any simple regression form of either occupancy, space mean
speed, or flow. The second module proposes two HOV schemes respectively
under lane utilization and travel time savings for speed equilibrium
between HOV and GP lanes. The schemes present distinct speed pairs by
congestion level, but speed of HOV lanes is identically ensured no less
than GP lanes under both schemes. The second module also covers an HOT
scheme that adopts value of time and value of reliability to formulate
HOT tolls with respect to speed of GP lanes. The third module identifies
lane management and congestion hot spots by contrasting the level of
service of managed lanes and GP lanes in a deterministic or stochastic
way. The case study indicates that lane management hot spots are
spatially and temporally dynamic, and a non-hot spot less likely turns
to a congestion hot spot without being a lane management hot spot as
transition, or vise versa. The last module develops two macroscopic
approaches to screening the policy combination set of managed lanes, and
eliminates the combinations by 60% in the selected scenario. Finally,
the optimal/non-inferior policies for non-eliminated combinations are
verified by solving such a case as a multi-objective binary integer
linear programming problem.
