Prediction Equilibrium for Dynamic Network Flows

We study a dynamic traffic assignment model, where agents base their instantaneous rout-ing decisions on real-time delay predictions. We formulate a mathematically concise model and define dynamic prediction equilibrium (DPE) in which no agent can at any point dur-ing their journey improve their predicted travel time by switching to a different route. We demonstrate the versatility of our framework by showing that it subsumes the well-known full information and instantaneous information models, in addition to admitting further realistic predictors as special cases. We then proceed to derive properties of the predictors that ensure a dynamic prediction equilibrium exists. Additionally, we define epsilon-approximate DPE wherein no agent can improve their predicted travel time by more than epsilon and pro -vide further conditions of the predictors under which such an approximate equilibrium can be computed. Finally, we complement our theoretical analysis by an experimental study, in which we systematically compare the induced average travel times of different predic-tors, including two machine-learning based models trained on data gained from previously computed approximate equilibrium flows, both on synthetic and real world road networks.

Automated summary

We propose a dynamic traffic assignment model where agents make instant routing decisions based on real-time delay predictions. The model encompasses full information and instantaneous information models, while allowing for more realistic predictors. We derive properties of the predictors that ensure the existence of a dynamic prediction equilibrium and define epsilon-approximate DPE. Experimental comparisons of different predictors' average travel times are conducted.