Fosgerau's travel time reliability ratio and the Burr distribution

作者：Taylor Michael A P^*

来源：Transportation Research Part B: Methodological , 2017, 97: 50-63.

DOI：10.1016/j.ijindorg.2016.12.001

• Travel time reliability
• Value of travel time reliability
• Reliability metrics
• Economic analysis
• Maximum likelihood
• Burr distribution

Recent international research has seen the development of methods for the inclusion of travel time reliability as a separate factor in economic analysis of transportation projects, including the valuation of travel time variability. Fosgerau's valuation method includes the consideration of travel time reliability in cost-benefit analysis by adding travel time variability to the set of generalised travel costs. This requires: (1) a defined unit of measurement for travel time variability, (2) estimates of the quantity of travel time variability, and (3) identification of the cost to travellers per unit of travel time variability. The chosen unit of measurement is the standard deviation of the travel time distribution, and the value of this unit of measurement can be defined relative to the average value of travel time by a reliability ratio that depends on user preference parameters (related to risk aversion) and the shape of the upper tail of the cumulative distribution cdf) of the travel time distribution. This shape is represented by a definite integral of the inverse of the cdf. Determining the shape of the cdf and its inverse function is facilitated if the distribution can be defined by an explicit algebraic function. The Burr (type XII) distribution is one distribution with this feature, and has been used to successfully represent observed travel time data. This paper describes the Burr distribution, demonstrates that it can provide a good representation of observed travel time data, and explains how it can be used to develop an exact expression for the reliability ratio and thus can aid the use of the method for the valuation of travel time reliability.

• 出版日期2017-3