Stochastic modeling of vehicle delay at two-lane highway work zones
The purpose of this research is to formulate analytical expressions for estimating expected vehicle delay as a function of two-lane highway work zone operating conditions. To capture the stochastic nature of work zone operations, initial effort has employed statistical estimation and Monte-Carlo simulation to identify distributions of average delay. Approximate analysis techniques were then used to compute user-specified percentile values of average delay per operating sequence per direction of travel. However, the use of Monte-Carlo simulation to simply generate random variables does not accurately characterize one-way traffic control. Principles of queueing theory were used to derive models for estimating expected delay. The models capture stochastic affects on expected delay as a function of directional demands, work zone physical length and observable traffic stream measures prevailing at work zones. These stochastic models hold underlying assumption that system operates under steady state condition. To evaluate validity of applying stochastic queueing model for delay prediction at two-lane highway work zone, the steady state assumption was assessed using a microscopic simulation model. The simulation model was used to evaluate the accuracy of the proposed stochastic models. The proposed delay expressions appear to adequately reflect the impacts of the work zone operation based upon the outcomes from the simulation model.
Cassidy, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our