As briefly described in Route Optimization Overview, a basic request consists of Model, Shipments, and Vehicles as required entities:
- Model captures settings and constraints for the entire request,
inclusive of both
Shipments
andVehicles
. - Shipments represent tasks or actual shipments that include pickup and
delivery
VisitRequest
s. Shipments have local settings and constraints. - Vehicles represent vehicles, drivers, or personnel. Vehicles also have local settings and constraints.
Each entity's properties describe part of an optimization problem at a particular level of granularity. Model-wide constraints are applied to all shipments and vehicles, while constraints and properties specified on shipments or vehicles are specific to a single shipment or vehicle.
For complete documentation on each message type, see the reference documentation
for ShipmentModel
(REST, gRPC), Shipment
(REST, gRPC),
and Vehicle
(REST, gRPC) messages.
OptimizeToursRequest
properties
Some commonly used properties of the top-level OptimizeToursRequest
message
(REST, gRPC) include the following:
searchMode
indicates whether to return the first solution that satisfies specified constraints or find the best possible solution within a set deadline.considerRoadTraffic
determines whether or not live traffic is used for routing and ETA estimation.populateTransitionPolylines
determines whether or not route polylines and route tokens are returned in the response.
Model properties
Some commonly used properties of the ShipmentModel
message (REST,
gRPC) include:
globalStartTime
represents the earliest start time of routes across all vehicles and shipments. No vehicle may start its first transitions and shipments before this time.globalEndTime
represents the latest end time of routes across all vehicles and shipments. All assigned shipments and transitions must be complete before this time.
Shipment properties
Some commonly used properties of the Shipment
message (REST, gRPC)
include:
pickups[]
anddeliveries[]
represent where a shipment can be picked up or dropped off.pickups[]
anddeliveries[]
properties both use theVisitRequest
message (REST, gRPC).loadDemands
represent the load required for a vehicle to complete a shipment. Vehicles' correspondingload_limits
(REST, gRPC) property represents how much load a vehicle can accommodate at one time. Read more about load in Load Demands and Limits.penalty_cost
represents the cost incurred if a shipment is skipped. Read more on costs in Cost Model Parameters.
Vehicle properties
Some commonly used properties of the Vehicle
message (REST, gRPC)
include:
startLocation
represents where a vehicle must start its route. This property is optional. If not specified, the vehicle's route starts at the location of its first assigned shipment.endLocation
represents where a vehicle must end its route. This property is optional. If not specified, the vehicle's route ends at the location of its last assigned shipment.startTimeWindows[]
represents when a vehicle can start its route. This property is optional.endTimeWindows[]
represents when a vehicle can start and end its route. Both properties are optional.loadLimits
represent the vehicle's capacity available to meet shipments' load demands. Read more about load demands and limits in Load Demands and Limits.
A complete example request in JSON format looks like:
{
"model": {
"shipments": [
{
"pickups": [
{
"arrivalLocation": {
"latitude": 37.73881799999999,
"longitude": -122.4161
}
}
],
"deliveries": [
{
"arrivalLocation": {
"latitude": 37.79581,
"longitude": -122.4218856
}
}
]
}
],
"vehicles": [
{
"startLocation": {
"latitude": 37.73881799999999,
"longitude": -122.4161
},
"endLocation": {
"latitude": 37.73881799999999,
"longitude": -122.4161
},
"costPerKilometer": 1.0
}
],
"globalStartTime": "2024-02-13T00:00:00.000Z",
"globalEndTime": "2024-02-14T06:00:00.000Z"
}
}
OptimizeTours
and BatchOptimizeTours
both consume request messages like the
example above, but in different ways. Before making a Route Optimization
request, it's important to understand the difference between the two methods: