A method for structuring stochastic travel time by using risk premiums of stochastic link flow

This study proposes a method for structuring stochastic link travel times in a road network with stochastic traffic demands. In our proposed method, the uncertainty of travel time is evaluated by link-flow-based indices. The mean link travel time is represented by using both its mean link flow and the corresponding risk premium of the stochastic link flow. The risk premium is defined by a concept of certainty equivalent derived from the relationship between a stochastic link flow and a link cost function. By expanding the concept of risk premium, a calculation method of link travel time covariance is proposed. Our proposed method defines bi-variate risk premiums, by which the mean of the product of two stochastic link travel times can be calculated. As a numerical calculation, we demonstrate our proposed method in a test network. Finally, we conclude and show future directions for evaluating real road networks