Verso’s API for route optimization has just been enhanced. We can now model pickups and deliveries in combination, which allows us to deal with a greater variety of situations. As always, the latest update is based on a new version of VROOM, the open-source software we work with to optimize itineraries.
A New Feature
The previous update allowed us to deal with objects that all had to be either loaded when the vehicle set out or brought back at the end of a round. In the trade, pickup and delivery is the phrase used to designate an object picked up and later delivered in the course of the same round. The newly added modeling entityshipment deals with this exact situation.
And all kinds of combinations are possible, for shipment objects can be combined with the previously managed job objects of single deliveries or pickups.
New Uses
It is now possible to optimize the deliveries and pickups of multiple clients to or from a variety of storage areas (warehouses, stores, etc.), all while taking into account the usual factors: priorities, load capacities, schedules, expertise in handling, and so on. When dealing with a multi-site fleet of vehicles, the optimization tool minimizes operating costs by determining the vehicles best suited for a series of pickups and its associated deliveries.
The new feature can also be applied to the transport of people, with possibilities for optimization combined with innovative solutions for carpooling or the use of common-carrier vehicles. In the case of individual requests for transportation, calculation time is so short that schedules can be updated and optimized in real time.
The Same Excellent Results as Always
As has consistently been the case, the results we achieve are both verifiable and reproducible. According to the well-established benchmark Li & Lim 100, our approach obtains absolute optimal results more than 46% of the time, for an average of just 1.22% over optimal cost. And everything is done in record time, with problems involving a hundred different locations solved on average in 256 milliseconds.
If you are interested in trying out this new functionality, contact us for an API key.
