In this paper, we propose the reliability premium as a generic concept that eliminates the utility difference between the expected utility and the utility of a travel choice that is subject to randomness in travel time–without specifying the underpinning utility function nor the travel choice domain to which it relates. Mathematically, the reliability premium quantifies the buffer or additional time a traveller is willing to pay beyond the expected outcome of a travel choice to eliminate the extra disutility due to travel time variability (TTV), thereby conceptualising the cost of TTV directly and intuitively in time units. We then discuss the reliability premium under, first, the Bernoulli approach, which focuses on route choice only, and second, the scheduling delay approach, which encompasses both departure or arrival time choice and route choice. Under the Bernoulli approach, we show that it is convenient to derive the monetary cost of travel time variability based on the reliability premium. In addition, we discuss the preservation of first-order and second-order stochastic dominance (SD) of the reliability premium, which removes the computational concern of using the reliability premium in reliable path-finding problems or related assignment models. Under the schedule delay framework, we derive formulations of the reliability premium for different applications and show the detailed impact of TTV on the resulting valuations. We find that the reliability premium can be effective in capturing the asymmetry and distributional tail of travel times for quantifying the TTV cost, especially for risk-averse users, making it suitable for evaluating the impact of TTV on travellers’ route choice decisions. Numerical examples are employed to elucidate the concept of the reliability premium and illustrate its practical application.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)