A stochastic approach to traffic congestion costs

The real world is a complex dynamic and stochastic environment. This is especially true for the traffic moving daily on our roads. As such, accurate modeling that correctly considers the real-world dynamics and the inherent stochasticity is very important, especially if government will base its road tax decisions on the outcomes of these models. The contemporary traffic prices, if any, however, do not reflect the external congestion costs. In order to induce road users to make the correct decision, marginal external costs should be internalized. To assess these costs, the public sector managers need accurate operational models. We show in this article that using a better representation and characterization of the road traffic, via stochastic queueing models, leads to a more adequate reflection of the congestion costs involved. Using extensive numerical experiments, we show the superiority of the stochastic traffic flow models.