Data describing planned routes and driven routes from a last-mile delivery vehicle routing problem

This dataset provides real-world last-mile delivery routes from a logistics company, offering valuable resources for research in route optimization and logistics. The dataset addresses a substantial gap between theoretical route planning and routing performed by drivers who often do not follow the proposed routes. The dataset comprises 19,497 pairs of routes, including both planned routes, generated by human dispatchers employing heuristic optimization tools, and the routes that the drivers eventually followed in practice. The routes were completed by 400 unique drivers, covering 10,864 unique locations. In the dataset, the routes are presented as sequences of stops, where each stop is a separate row in the presented datafile. The dataset includes detailed information on routes, stops and drivers' identifiers, timing, and distance measurements. It allows for a deeper analysis of driver behavior patterns, factors of route deviations, and their impact on delivery performance.