Network effects of percentile-based route choice behavior for stochastic travel times under exogenous capacity variations

This study analyzes the properties of network reliability and efficiency with regard to the role of route choices under stochastic travel times, here due to exogenous link capacity variations. A percentile-based route choice model is derived, which has as underlying behavioral assumption that drivers make routing decisions based on route travel time distributions collected from past experiences and do so by accounting for the travel time budget that needs to be allocated in order to ensure a desired probability of avoiding the trip taking longer than the allocated travel time budget. This may represent, e.g., the routing decisions of commuters without real-time traffic information. The model application to the Sioux Falls road network then shows how variations in routing percentiles of homogeneous and segmented driver populations substantially impact the route travel time distributions as well as network performance metrics, such as buffer times and delay times, and how these effects vary with travel demand.