To re-route, or not to re-route: Impact of real-time re-routing in urban road networks

Route planning represents a major challenge with substantial impact on safety, economy, and even climate. An ever-growing urban population caused significant increase in commuting times, therefore, stressing the prominence of efficient real-time route planning. In essence, goal is to compute fastest reach target location realistic environment where traffic conditions are time-evolving. Consequently, large volume data potentially required continuously updated. We thereby address re-routing problem answer questions such as when, how often, worthwhile. base our study real dataset, comprising travel times road segments New York, London, Chicago, collected over three months. By exploiting this we implement an optimal algorithm, able mimic ideal predictions segment speeds network. Thereby, allowing us lower bound travel-time serve reference against other routing techniques. Mainly, quantify achieved gain static, no re-routing, continuous strategies. Surprisingly, find that sufficiently stable for short time windows, vehicle very seldom useful when accurate statistics at departure time. Typically, should only be triggered during rush hours, long routes, passing through well-identified segments.