The Department of Transport (DoT) commissioned a study of the effect of highway schemes on journey times and their variability. Some findings are presented of the preliminary phase of this research, which aimed to review existing methodologies, formulate a preferred approach, and apply the methodology to a selection of schemes. Sources of variability of travel time include variation by season, day of week, or time of day, `random' variation, and catastrophic delays due to incidents. There are two possible approaches for estimating the day-to-day variability for a specific journey: (1) using aggregate variables, such as mean travel time or average speed, that express characteristics of a complete complete trip; and (2) constructing the estimate in terms of variability on individual links and junctions on the route. The second approach was used here, to obtain a model adaptable to different types of road network. Submodels are established for the standard deviation of travel time at various types of link and junction, and for incidents. The submodels were calibrated against survey data for London and Edinburgh, and applied to three road schemes. Procedures were established to evaluate travel time variability, and incorporated into a computer program, VARTIME. More research is needed to allow for variable junction traffic conditions.
Abstract