Incorrect driving time estimate when using curbside

Directions API Route Optimization API
1

Hi,

We’ve found an interesting case where providing curbside causes the driving_time to increase a lot higher than expected.

Swapping the curbside of fCqIQFciILD1eilVU2Qm from any to left causes the driving time to be increased by ~30 minutes, even though the driving distance only increases by ~100m.

Here is the optimization request:

{
  "configuration": {
    "routing": {
      "calc_points": true,
      "consider_traffic": false,
      "fail_fast": true,
      "network_data_provider": "openstreetmap",
      "return_snapped_waypoints": true,
      "snap_preventions": [
        "motorway",
        "tunnel",
        "ferry"
      ]
    }
  },
  "objectives": [
    {
      "type": "min",
      "value": "completion_time"
    }
  ],
  "relations": [],
  "services": [
    {
      "address": {
        "curbside": "any",
        "lat": -33.9108078,
        "location_id": "hdmMM9xdPoOmyYZrIQQQ",
        "lon": 151.210465,
        "name": "340/7 Letitia Street",
        "street_hint": "340/7 Letitia St"
      },
      "allowed_vehicles": [],
      "duration": 360,
      "id": "hdmMM9xdPoOmyYZrIQQQ",
      "name": "340/7 Letitia Street",
      "priority": 2,
      "time_windows": []
    },
    {
      "address": {
        "curbside": "left",
        "lat": -33.9095491,
        "location_id": "fCqIQFciILD1eilVU2Qm",
        "lon": 151.2119761,
        "name": "629/5 Bindon place",
        "street_hint": "629/5 Bindon Pl"
      },
      "allowed_vehicles": [],
      "duration": 360,
      "id": "fCqIQFciILD1eilVU2Qm",
      "name": "629/5 Bindon place",
      "priority": 2,
      "time_windows": []
    }
  ],
  "shipments": [],
  "vehicle_types": [
    {
      "profile": "car",
      "service_time_factor": 1,
      "speed_factor": 1,
      "type_id": "vhc_iiZd8BWoOVhdotLKz08Tdc4tjv92"
    }
  ],
  "vehicles": [
    {
      "earliest_start": 1724888289,
      "return_to_depot": false,
      "start_address": {
        "curbside": "any",
        "lat": -33.9358304,
        "location_id": "1yxunLKL",
        "lon": 151.2078147
      },
      "type_id": "vhc_iiZd8BWoOVhdotLKz08Tdc4tjv92",
      "vehicle_id": "iiZd8BWoOVhdotLKz08Tdc4tjv92"
    }
  ]
}
2

Thanks for reporting this. We already noticed this and the fix is on the way.

3

Hi @easbar ,

Just checking in to see if there’s been any progress on this bug? We’re still seeing it happen.

Example request:

{
  "configuration": {
    "routing": {
      "calc_points": true,
      "consider_traffic": false,
      "fail_fast": false,
      "network_data_provider": "openstreetmap",
      "return_snapped_waypoints": true,
      "snap_preventions": [
        "motorway",
        "tunnel",
        "ferry"
      ]
    }
  },
  "objectives": [
    {
      "type": "min",
      "value": "completion_time"
    }
  ],
  "relations": [],
  "services": [
    {
      "address": {
        "lat": -34.0216022,
        "location_id": "1yguaG2oSwpnTq5eJe7z",
        "lon": 151.010344,
        "name": "4/10 Owen Jones Row",
        "street_hint": "Owen Jones Row"
      },
      "duration": 120,
      "group": "medium",
      "id": "1yguaG2oSwpnTq5eJe7z",
      "name": "4/10 Owen Jones Row",
      "priority": 2,
      "size": [
        1
      ],
      "time_windows": []
    }
  ],
  "shipments": [],
  "vehicle_types": [
    {
      "capacity": [
        2147483647
      ],
      "profile": "car_delivery",
      "service_time_factor": 1,
      "speed_factor": 1,
      "type_id": "vhc_2koj50miEgrSl5rCdil4"
    }
  ],
  "vehicles": [
    {
      "earliest_start": 1727211600,
      "min_jobs": 1,
      "return_to_depot": false,
      "start_address": {
        "curbside": "left",
        "lat": -34.0167836,
        "location_id": "Fnv85NaT5tbrnWrK10H1",
        "lon": 151.0218885,
        "name": "12 Parkwood Drive",
        "street_hint": "Parkwood Drive"
      },
      "type_id": "vhc_2koj50miEgrSl5rCdil4",
      "vehicle_id": "2koj50miEgrSl5rCdil4"
    }
  ]
}
4

Sorry, there was a regression with the new version and so we had to roll it back. Another version will be deployed in ~2 days will hopefully fix both problems.

5

Yes, I made sure your example was already fixed some days ago, but the problem happens again now, because like @karussell said we briefly went back to the older, broken version.