This guide demonstrates how the number of vehicles provided in a Route Optimization solution can vary depending on the request parameters.
The Route Optimization API not only optimizes shipment completion order, but it also assigns those shipments to vehicles in order to optimize costs under constraints you manage.
In the first example, the number of vehicles matches the number of shipments, with all vehicles sharing the same cost and location properties. Each vehicle has a cost per operating hour and cost per traveled kilometer, which will help to minimize travel time and distance. One might expect multiple vehicles to be assigned shipments, but the example response shows the least cost solution given the specified cost model parameters.
See an example request with multiple vehicles
{ "model": { "globalStartTime": "2023-01-13T16:00:00-08:00", "globalEndTime": "2023-01-14T16:00:00-08:00", "shipments": [ { "deliveries": [ { "arrivalLocation": { "latitude": 37.789456, "longitude": -122.390192 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 100.0 }, { "deliveries": [ { "arrivalLocation": { "latitude": 37.789116, "longitude": -122.395080 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 5.0 }, { "deliveries": [ { "arrivalLocation": { "latitude": 37.795242, "longitude": -122.399347 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 50.0 } ], "vehicles": [ { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerHour": 50.0, "costPerKilometer": 10.0 }, { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerHour": 50.0, "costPerKilometer": 10.0 }, { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerHour": 50.0, "costPerKilometer": 10.0 } ] } }
See a response to the request with multiple vehicles
{ "routes": [ { "vehicleStartTime": "2023-01-14T00:00:00Z", "vehicleEndTime": "2023-01-14T00:28:22Z", "visits": [ { "isPickup": true, "startTime": "2023-01-14T00:00:00Z", "detour": "0s" }, { "shipmentIndex": 2, "isPickup": true, "startTime": "2023-01-14T00:02:30Z", "detour": "150s" }, { "startTime": "2023-01-14T00:08:55Z", "detour": "150s" }, { "shipmentIndex": 2, "startTime": "2023-01-14T00:21:21Z", "detour": "572s" } ], "transitions": [ { "travelDuration": "0s", "waitDuration": "0s", "totalDuration": "0s", "startTime": "2023-01-14T00:00:00Z" }, { "travelDuration": "0s", "waitDuration": "0s", "totalDuration": "0s", "startTime": "2023-01-14T00:02:30Z" }, { "travelDuration": "235s", "travelDistanceMeters": 795, "waitDuration": "0s", "totalDuration": "235s", "startTime": "2023-01-14T00:05:00Z" }, { "travelDuration": "496s", "travelDistanceMeters": 1893, "waitDuration": "0s", "totalDuration": "496s", "startTime": "2023-01-14T00:13:05Z" }, { "travelDuration": "171s", "travelDistanceMeters": 665, "waitDuration": "0s", "totalDuration": "171s", "startTime": "2023-01-14T00:25:31Z" } ], "metrics": { "performedShipmentCount": 2, "travelDuration": "902s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "800s", "totalDuration": "1702s", "travelDistanceMeters": 3353 }, "routeCosts": { "model.vehicles.cost_per_kilometer": 33.53, "model.vehicles.cost_per_hour": 23.638888888888889 }, "routeTotalCost": 57.168888888888887 }, { "vehicleIndex": 1 }, { "vehicleIndex": 2 } ], "skippedShipments": [ { "index": 1 } ], "metrics": { "aggregatedRouteMetrics": { "performedShipmentCount": 2, "travelDuration": "902s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "800s", "totalDuration": "1702s", "travelDistanceMeters": 3353 }, "usedVehicleCount": 1, "earliestVehicleStartTime": "2023-01-14T00:00:00Z", "latestVehicleEndTime": "2023-01-14T00:28:22Z", "totalCost": 62.168888888888887, "costs": { "model.vehicles.cost_per_hour": 23.638888888888889, "model.shipments.penalty_cost": 5, "model.vehicles.cost_per_kilometer": 33.53 } } }
The solver assigns all shipments to only one vehicle, skipping one shipment
despite ample vehicle availability. This is because the cost of operating
additional vehicles is too high to justify, and it isn't cost effective for
any vehicle to complete the skipped shipment given its low penalty cost.
Despite available vehicle capacity, one vehicle can perform all assigned
shipments in the most cost-effective manner. The vehicles in the request don't
have the usedIfRouteIsEmpty property set (see the Vehicle message
documentation (REST, gRPC) for details), so they incur no cost if
unused.
Changing the cost parameters to prioritize globally shorter solutions rather
than individually shorter vehicle routes causes more vehicles to participate
in the solution. The next example request replaces Vehicle.costPerHour with
the global ShipmentModel.globalDurationCostPerHour, prioritizing solutions
that are shorter in total over operating time for any given vehicle. The penalty
cost for shipment[1] is also increased to reduce its likelihood of being
skipped.
See an example request using
globalDurationCostPerHour
{ "model": { "globalStartTime": "2023-01-13T16:00:00-08:00", "globalEndTime": "2023-01-14T16:00:00-08:00", "globalDurationCostPerHour": 150.0, "shipments": [ { "deliveries": [ { "arrivalLocation": { "latitude": 37.789456, "longitude": -122.390192 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 100.0 }, { "deliveries": [ { "arrivalLocation": { "latitude": 37.789116, "longitude": -122.395080 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 75.0 }, { "deliveries": [ { "arrivalLocation": { "latitude": 37.795242, "longitude": -122.399347 }, "duration": "250s" } ], "pickups": [ { "arrivalLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "duration": "150s" } ], "penaltyCost": 50.0 } ], "vehicles": [ { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerKilometer": 10.0 }, { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerKilometer": 10.0 }, { "endLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "startLocation": { "latitude": 37.794465, "longitude": -122.394839 }, "costPerKilometer": 10.0 } ] } }
The result shows that using the global cost per hour cost parameter results in all three vehicles being used instead of just one.
See a response to the request using
globalDurationCostPerHour
{ "routes": [ { "vehicleStartTime": "2023-01-14T00:00:00Z", "vehicleEndTime": "2023-01-14T00:16:20Z", "visits": [ { "shipmentIndex": 2, "isPickup": true, "startTime": "2023-01-14T00:00:00Z", "detour": "0s" }, { "shipmentIndex": 2, "startTime": "2023-01-14T00:09:19Z", "detour": "0s" } ], "transitions": [ { "travelDuration": "0s", "waitDuration": "0s", "totalDuration": "0s", "startTime": "2023-01-14T00:00:00Z" }, { "travelDuration": "409s", "travelDistanceMeters": 1371, "waitDuration": "0s", "totalDuration": "409s", "startTime": "2023-01-14T00:02:30Z" }, { "travelDuration": "171s", "travelDistanceMeters": 665, "waitDuration": "0s", "totalDuration": "171s", "startTime": "2023-01-14T00:13:29Z" } ], "metrics": { "performedShipmentCount": 1, "travelDuration": "580s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "400s", "totalDuration": "980s", "travelDistanceMeters": 2036 }, "routeCosts": { "model.vehicles.cost_per_kilometer": 20.36 }, "routeTotalCost": 20.36 }, { "vehicleIndex": 1, "vehicleStartTime": "2023-01-14T00:00:00Z", "vehicleEndTime": "2023-01-14T00:18:54Z", "visits": [ { "shipmentIndex": 1, "isPickup": true, "startTime": "2023-01-14T00:00:00Z", "detour": "0s" }, { "shipmentIndex": 1, "startTime": "2023-01-14T00:08:24Z", "detour": "0s" } ], "transitions": [ { "travelDuration": "0s", "waitDuration": "0s", "totalDuration": "0s", "startTime": "2023-01-14T00:00:00Z" }, { "travelDuration": "354s", "travelDistanceMeters": 1192, "waitDuration": "0s", "totalDuration": "354s", "startTime": "2023-01-14T00:02:30Z" }, { "travelDuration": "380s", "travelDistanceMeters": 1190, "waitDuration": "0s", "totalDuration": "380s", "startTime": "2023-01-14T00:12:34Z" } ], "metrics": { "performedShipmentCount": 1, "travelDuration": "734s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "400s", "totalDuration": "1134s", "travelDistanceMeters": 2382 }, "routeCosts": { "model.vehicles.cost_per_kilometer": 23.82 }, "routeTotalCost": 23.82 }, { "vehicleIndex": 2, "vehicleStartTime": "2023-01-14T00:00:00Z", "vehicleEndTime": "2023-01-14T00:16:14Z", "visits": [ { "isPickup": true, "startTime": "2023-01-14T00:00:00Z", "detour": "0s" }, { "startTime": "2023-01-14T00:06:25Z", "detour": "0s" } ], "transitions": [ { "travelDuration": "0s", "waitDuration": "0s", "totalDuration": "0s", "startTime": "2023-01-14T00:00:00Z" }, { "travelDuration": "235s", "travelDistanceMeters": 795, "waitDuration": "0s", "totalDuration": "235s", "startTime": "2023-01-14T00:02:30Z" }, { "travelDuration": "339s", "travelDistanceMeters": 1276, "waitDuration": "0s", "totalDuration": "339s", "startTime": "2023-01-14T00:10:35Z" } ], "metrics": { "performedShipmentCount": 1, "travelDuration": "574s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "400s", "totalDuration": "974s", "travelDistanceMeters": 2071 }, "routeCosts": { "model.vehicles.cost_per_kilometer": 20.71 }, "routeTotalCost": 20.71 } ], "metrics": { "aggregatedRouteMetrics": { "performedShipmentCount": 3, "travelDuration": "1888s", "waitDuration": "0s", "delayDuration": "0s", "breakDuration": "0s", "visitDuration": "1200s", "totalDuration": "3088s", "travelDistanceMeters": 6489 }, "usedVehicleCount": 3, "earliestVehicleStartTime": "2023-01-14T00:00:00Z", "latestVehicleEndTime": "2023-01-14T00:18:54Z", "totalCost": 112.14, "costs": { "model.vehicles.cost_per_kilometer": 64.89, "model.global_duration_cost_per_hour": 47.25 } } }
In this response, all three vehicles are in use (per metrics.usedVehicleCount)
with each vehicle being assigned one shipment to complete. With identical start
locations, end locations, and costPerKilometer, all three vehicles are
effectively interchangeable, so it doesn't matter which shipment is assigned to
which vehicle.
The globalDurationCostPerHour causes the optimizer to find a solution that is
shorter overall: the difference between earliestVehicleStartTime and
latestVehicleEndTime is only 18 minutes 54 seconds as opposed to the 28
minutes and 22 seconds in the previous response. That said, the
metrics.costs.model.vehicles.cost_per_kilometer has increased, reflecting more
total traveled distance by the three used vehicles. This demonstrates one way in
which the cost model lets you make trade-offs:
- Increased global time cost: Increase vehicle utilization to minimize overall completion time, at the cost of more vehicle distance and time spent in transit.
- Increased vehicle time cost: Decrease vehicle utilization and time spent in transit, at the cost of a longer overall solution.
Note that the globalDurationCostPerHour value of 150.0 in this example is set
at three times the individual vehicles' costPerHour of 50.0 from the previous
example. This global cost value effectively expects that all three vehicles will
operate simultaneously, but in practical settings such assumptions may not
reflect reality and may in fact have a negative effect on result quality.
As described in Cost Model Parameters, all cost parameters are expressed in the same dimensionless units but can have very different meanings. Typically, cost model parameter values should be grounded in reality as much as possible, as artificial costs like those in this example may cause the API to optimize for objectives that don't match your intent.