The usefulness or uselessness of real-time itinerary calculation in road networks

This paper investigates the impact of real-time traffic data on route planning efficiency. By comparing different strategies, from static travel times to real-time updates, the study quantifies the benefits and computational costs of dynamic routing. Using real-world datasets from cities like London and NYC, the authors evaluate the trade-offs between accuracy and performance. Future work will focus on optimizing the use of real-time data by identifying the most critical moments and locations for route recalculations.

Route planning has become a major challenge, with a significant impact on the economy, safety and the climate. It involves providing each user with a route offering the shortest travel time, even if traffic conditions change. Thus, such a strategy requires reconsideration of the route to be taken on an ongoing basis, as conditions evolve. However, taking such real-time data into account has a high impact on the computational resources required. We therefore quantify here the gain brought by real-time data. We compare routes obtained using statistical data, versus real-time. We also provide a lower bound on travel time, with an algorithm that would be able to predict the future perfectly. Our results, based on a real data set, surprisingly show that real-time is actually of little use.