Drone Delivery Routing
Drone delivery offers the promise of quick and efficient package transportation, potentially reducing road traffic and carbon emissions. We developed an algorithm and an interactive tool to minimize the total delivery time by coordinating the trajectory of a truck and multiple drones.
The tool allows changing the drone features s such as range, speed, and fleet size. Follow the instructions on the left. Subsequent versions of the tool will include multiple trucks, multi-drone launching, hovering restrictions, no-flight zones, 3D drone navigation, and onboard charging.
Vehicle Routing Problem
The Vehicle Routing Problem is to find the most efficient routes for a fleet of vehicles to deliver goods or services to a set of destinations, typically with constraints like vehicle capacity.
The problem was first posed by George Dantzig and John Ramser in 1959 in their paper "The Truck Dispatching Problem," published in "Management Science." Since then and according to Google Scholar, more than a million papers have been dedicated to solving this problem given its complex combinatorial nature.
Despite the large body of literature, practical applications of the problem on real networks are rare. The following tool demonstrates our front-end algorithm for the vehicle routing problem. You can click on the map to add points, move them by drag/dropping, and change the fleet size and capacity of the vehicles.
Reach out to us if interested in testing it out.
Two-Echelon Vehicle Routing
The Two-Echelon Vehicle Routing Problem is a logistics optimization challenge of transporting goods from a central depot to customers through intermediate satellites in two stages: first using larger vehicles to satellites and then smaller vehicles for final recipients. The latter is also known as the last mile.
The two modes of transportation can be picked from any of driving, cycling, walking, or even transit depending on the application. An example would be when a delivery driver finds a parking space to make a collection of deliveries within a certain proximity on foot, choosing driving as the first and walking as the second mode.
We developed a two-echelon vehicle routing algorithm and optimized it for speed and efficiency in terms of minimizing total delivery time within a reasonable computation time using real routing data. The tool allows for picking the two modes of transportation, and the capacity and size of the second mode’s fleet.
Generative VRP
Last-mile logistics planning requires measuring potential delivery costs to assess whether a particular market is profitable. VRP-Gen is a web-based vehicle route generator that takes a service area as input, randomly generates customers uniformly or based on customer densities, solves variants of VRP, and tracks system performance metrics such as total delivery time.