Toward an automated methodology for the valuation of reliability

Abstract

There is a significant body of research related to the valuation of reliability in transportation. Such work has tended to rely on the stated preference methodology, where respondents are asked to make trade-offs between the mean and standard deviation of travel time. The literature has suggested that a revealed preference (RP) methodology may provide an alternative means of estimating a value of reliability. In this article we show how emerging data sources reveal travelers'preferences and, in combination with traditional choice modeling methods, can be used to estimate a value of reliability. We illustrate this RP methodology using smart-card data from the multimodal public transport network of London, UK. We are able to estimate a useable "reliability ratio" for three of four public transport modes modeled, and account for an unexpected result based upon the data available.