Calibration and Validation of a Microsimulation Model of Lane Closures on a Two-Lane Highway Work Zone

Abstract

Lane closure on a two-lane work zone can negatively affect operational performance. Although a microsimulation model is often used to quantify and predict these negative impacts, it requires proper calibration to produce meaningful results. The sixth edition of the Highway Capacity Manual (HCM-6) proposed a microsimulation model using limited empirical data from King County, Washington. Based on the authors’ knowledge, this paper is the first to examine the applicability of the HCM-6 two-lane work zone model in a different geographical area. A preliminary application of the HCM-6 model to two test sites in Nebraska found that the HCM-6 estimation of the mean headways and travel times were statistically different from observed data at the 5% significance level. Therefore, this paper proposed an automatic microsimulation calibration methodology for identifying parameters that can match the distribution or the mean value of observed traffic data. To the best of the authors’ knowledge, this is the first paper to show that a two-lane work zone microsimulation model calibrated using the empirical distributions of traffic measures can replicate the corresponding distributions at the 5% significance level. However, if the model was calibrated using the mean value of traffic measures, it could not replicate empirical distributions. This model was also validated using field data not used for calibration. The methodology proposed in this paper will benefit those who design and operate work zones because traffic agencies are interested in the variety of traffic conditions, not just the average. This paper concludes with a discussion on model transferability.