Abstract:
These notes describe some sources of variability in vehicle travel times when traversing a link as part of an urban journey. The aim is to look at possible inter-relationships between these times on adjacent links, and to disentangle sources of variability regardless of the size of that variability, so that some sort of simulation from (conditionally) independent components becomes possible. The approach is a mix of theoretical and empirical. There are some suggestions as to how link travel times might be simulated parametrically.
The data-based analyses are based on measurements made by 'probe' vehicles that traversed a variety of routes in a suburban area of Chicago between 1pm and 7pm on many weekdays in the period June to August 1995. These data are described more fully in other project reports.
tr44.pdf