Problem model

In general a pragmatic problem is split into two required and one optional parts:

• plan (required) models a work to be performed by vehicles taking into account all related constraints, such as time windows, demand, skills, etc.
• fleet (required) models available resources defined by vehicle types.
• objectives (optional) defines objective functions as goal of whole optimization.

Modeling jobs

A work which has to be done is model by list of jobs defined in plan.

Check next job section for detailed explanation.

Modeling vehicles

Vehicles are defined by fleet.vehicles property which specifies array of vehicle types, not specific vehicles.

More details can be found in vehicle type section.

Relation between jobs and vehicles

An optional plan.relations property specifies relations between multiple jobs and single vehicle. It is useful to lock jobs to a specific vehicle in any or predefined order.

Check relations section for more details.

Job and vehicle constraints

There are multiple strict constraints that should be matched on jobs and vehicles.

Demand and capacity

Each job should have demand property which models a good size in abstract integral units:

            "demand": [
              1
            ]

It is required, but you can set demand to zero in case it is not needed. It can be multidimensional array.

A capacity property is a vehicle characteristic which constraints amount of jobs can be served by vehicle of specific type based on accumulated demand value. Total demand should not exceed capacity value.

Time windows

Optionally, each job can have one or more time window:

                "times": [
                  [
                    "2019-07-04T09:00:00Z",
                    "2019-07-04T18:00:00Z"
                  ],
                  [
                    "2019-07-05T09:00:00Z",
                    "2019-07-05T18:00:00Z"
                  ]
                ]

Time windows are strict: if no vehicle can visit a job in given time ranges, then the job is considered as unassigned.

Vehicle time is limited per each shift and has required start optional end time:

            "start": {
              "earliest": "2019-07-04T09:00:00Z",
              "location": {
                "lat": 52.5316,
                "lng": 13.3884
              }
            },
            "end": {
              "latest": "2019-07-04T18:00:00Z",
              "location": {
                "lat": 52.5316,
                "lng": 13.3884
              }
            }

More details about shift property can be found in vehicle type section.

Clustering

Some jobs can be clustered together to have more realistic ETA, check vicinity clustering section.