Traffic Priority Mechanisms

Abstract

In the future world of self-driving vehicles, intersections will be managed by computers that send individual commands to each passing vehicle. This paper proposes to make traffic priority contingent on the self-reported value of time of vehicle occupants. A model of two merging roads with stochastic traffic is developed. Algorithms for calculation of optimal exit sequences, accounting for time value heterogeneity, are characterized. Welfare costs of a limited planning horizon are assessed. An incentive-compatible scheme of payment for priority is calculated. Under mild conditions, the mechanism is Pareto improving, that is, benefits all parties. The optimality of the traffic volume and composition under the optimal exit regulation is established.