Practical Benefits Analysis of Travel Time Reliability from Automated Detection

Resources for roadway infrastructure capacity expansion remain scarce, competition for finances is ubiquitous, and the emphasis continues to shift to operations, management, and efficiency. These conditions point directly to economic benefit–cost analysis, yet economic appraisal of any sort continues to make only meager inroads into institutional decision making. Barriers to widespread practical application include a lack of understanding and inconsistent application. The concept and measure of travel time reliability is a quintessential example of those two barriers. This paper clarifies some of these issues. First is an exploration of how travel time reliability is measured and how it is valued for use in economic analysis. This has become increasingly prominent in just the past 10 years. One remarkable example from a ramp meter study is highlighted. To illustrate the issues further, this discussion is followed by an assessment of 15-min automated detector data during weekday peak periods over 3 years on a section of the Beltline freeway in Madison, Wisconsin. The assessment demonstrates the measure of reliability used does matter and reinforces the need for consistency in practice and for consistency in estimates of the value of reliability. Of those measures in use, this paper recommends standardizing the difference between 90th percentile and 50th percentile (median) travel time, pending further investigation into whether it is a particularly efficient measure. Last, the paper concludes with a suggestion for how to put a value on reliability in the absence of sufficient disaggregate data.