There is increasing interest in improving travel time reliability in New Zealand (NZ) and elsewhere, as traffic congestion increases and users find their trip times become less predictable. NZ has adopted a method for estimating the benefit of an improvement in trip time reliability and incorporating that benefit in the economic appraisal of transport projects. The method entails estimating the change in the standard deviation of trip time for each portion of a trip, and assumes that the travel time for each portion of a trip is statistically independent of the travel times for other portions of the trip. This paper describes a study of travel times on the 11.9 km long eastern section of the Central Circular Route, which is part of the Tokyo Metropolitan Expressway system. The study involved dividing the section of expressway into 39 ôlinksö and collecting the flow rates and speeds on each ôlink" on 93 days. Analysis of the data reveals that the travel time for each ôlinkö (or trip portion) is correlated with the travel times for other trip portions. It is shown that ignoring this correlation results in substantial over-estimation of the benefits of reducing the standard deviation of the travel time for individual trip portions. (a) For the covering record of the conference, please refer to ITRD no. E218380.
Abstract