Check out the the August 2022 (22.06) release of NVIDIA cuOpt.
New Features
-
Vehicle - Order Match - Provide options to include additional capacity and demand constraints that can be used to allow/disallow specific vehicles to order types. This can also be extended for dispatch optimization, task routing or value added services. User can specify the constraints in a Boolean matrix of size nVehicles x nOrders. For instance, if the matrix(v1, c1) is set to false, then vehicle v1 is not allowed to serve customer c1. The default setting for this matrix sets all entries to true. If multiple vehicles match to a specific customer or order, the solver selects the option that minimizes the stated cost objective. Example below:
Consider a case with three vehicles (0, 1, 2) and six orders (0, 1, 2, 3, 4, 5).
The constraint is such that orders 2 and 4 cannot be served by vehicle 1.
We would set the constraints the following way:order_vehicle_match = [[True]*3, [True]*3, [True, False, True], [True]*3, [True, False, True], [True]*3] vehicle_order_match = cudf.DataFrame(order_vehicel_match).T data_model.set_vehicle_order_match(vehicle_order_match) -
Multi-Cost Support for Mixed Fleet Routing - The use cases reflects a pick-up and drop-off problem with delivery windows using a mixed fleet. These can be relevant to food delivery where the fleet has mixed types such as EV, bikes, cars, trucks with each having its own set of constraints. It can also be applied to intra logistics where activities such as picking can be performed by manual labor with and without forklifts, cranes, cobots and robots. Find these in the cuOpt container and run them via:
‘docker run --rm --network=host --gpus=all nvcr.io/ea-reopt-member- zone/ea-cuopt:22.06 2 jupyter-notebook` Pdptw_mixed_fleet.ipynb - demonstrates heterogenous fleet modeling for a pickup and delivery problem with time windows. -
Add additional features to the cuopt server version to support:
• Multi depot
• Vehicle break
• Precedence constraint
We appreciate your feedback! Please post on the forum or email us at cuopt@nvidia.com.
best,
NVIDIA cuOpt Team