Relationship between mean and day-to-day variation in travel time in urban networks

The day-to-day reliability of transportation facilities significantly affects travel behavior. To better understand how travelers use these facilities, it is critical to understand and characterize this reliability for different facilities. Early work in this area assumed that the variance of day-to-day travel times (a measure of the inverse of reliability) increases proportionally with the mean travel time; i.e., as the mean travel time increases, travel time reliability decreases. However, recent empirical data for a single bottleneck facility and a small urban network suggest a more complex relationship that exhibits hysteresis. When this phenomenon is present, the variance in travel time is larger as the mean travel time decreases (congestion recovery) than as the mean travel time increases (congestion onset). This paper presents an elegant theoretical model to describe the variance of travel times across many days in an urban network. This formulation shows that the hysteresis behavior observed in empirical floating car data on urban networks should not be unexpected, and that it is linked to the hysteresis loops that often exist in the Macroscopic Fundamental Diagram of urban traffic. To verify the validity of this formulation, data from a micro-simulation of the City of Orlando, Florida, are V. V. Gayah (&) Department of Civil and Environmental Engineering, The Pennsylvania State University, 223A Sackett Building, University Park, PA 16802, USA e-mail: [email protected] V. V. Dixit School of Civil and Environmental Engineering, University of New South Wales, CE 106 Civil Engineering Building (H20), Sydney 2052, Australia e-mail: [email protected] S. I. Guler Institute for Transport Planning and Systems, ETH Zurich, Wolfgang-Pauli Strasse 15 HIL F41.2, Zurich 8093, Switzerland e-mail: [email protected] 123 EURO J Transp Logist (2014) 3:227–243 DOI 10.1007/s13676-013-0032-2