Introduction
The TD-R Matrix is a Simacan product where for all possible routes between known locations, we calculate the most realistic distance for each route and the expected driving time for each time block for each weekday.
The TD-R Matrix is used by third-party scheduling tools to generate more accurate and realistic schedules. The Matrix is supplied as a file/API. The abbreviation TD-R stands for:
- T > Time
- D > Distance
- R > Restrictions
What do we take into account
The route takes into account:
- The generated route is the expected fastest route.
- Permitted roads by vehicle type (normal car or normal truck).
- Road works with a longer duration (not ad hoc works).
The route does not take into account:
- The shortest route or a combination between a fast and short route.
- Approved roads for all vehicle types (e.g. trucks that deviate from standard sizes and bicycles).
- Ad hoc work or work of short duration.
- Time-bound road closures.
The driving time takes into account:
- Expected effects of the time and day of the week when the ride is driven at speed. For example, expected files.
- Effects of maximum speed and the maximum allowed speed per vehicle type (normal car or normal truck).
- Seasonal factors (to be added at extra cost), for example the difference in traffic jams between autumn and spring or less traffic congestion expected during holiday periods.
Driving time does not take into account:
- Maneuver Attractions around a location. This is location dependent and is seen as part of the location.
- An extra buffer for the chance that a delay lasts longer than expected.
- The TD Matrix takes driving time into account, not travel time. Matters such as break times, refueling, missing a turn, having to search for a location, etc. are not included in this.
The expected driving time takes into account the fastest route and the expected crowds at any time of every day of the week are taken into account. The current benchmarks of the TD Matrix take into account the less busy roads due to Corona.
What we deliver
Simacan delivers per start and end location, the distances and the driving time. The (delay) driving time is per day of the week and per 30 minutes. The final driving time is calculated as: freeflow driving time + (delay driving time * deceleration factor). The routes are for the active locations in Simacan Masterdata including the first and last kilometers.
We provide 3 files:
- timedistances.csv
- delay_monday.csv (a different CSV for each day of the week
- delay_factors_20210812.csv (a different CSV every day) (Seasonal Factors)
timedistances.csv
The distance and freeflow travel time (time without congestion).
Each row is an origin-destination pair. The row contains columns with the distances (meters) and the freeflow travel time (seconds).
delay_monday.csv
The delay travel time for a specified time.
Each row is an origin-destination pair.
We have the delays in CSV format and parquet format.
The parquet file has columns with delays per 5 minutes of that specific day of the week (ie 256 columns). The delay is in seconds. The first column is the travel time from 00:00. The last 6 columns are the times of the next day. So in the case of the parquet file from Monday, these are 00:00 to 00:25 on Tuesday.
The CSV files have columns of delays per 30 minutes. It is not every 5 minutes, because the file will then be 6 times larger. The first column delay_0 is the travel time for 00:12:30. This is the average travel time for the first 6 columns of the parquet file. So the travel time between 00:00, 00:05, 00:10, 00:15, 00:20, 00:25.
delay_factors_20210812.csv
More/less delay expected on a specific day for a specific region.
During holidays, for example, there is less delay.
Each row is an origin-destination pair. The row contains a column with the delay factor.
To take into account when implementing the TD-R Matrix
Consequences of unexpected congestion
The TD matrix consists of the driving time that has the best chance of driving. This means that there is always a chance that driving time will be longer if unexpected congestion occurs. If there are more traffic jams than expected at the start of the journey, this will cause the vehicle to arrive later at all upcoming stops. Because the TD array has no extra buffer in case this happens.
However, the parties do not want the driving time that has the best chance, but a driving time that is on average the most accurate.
Predict driving times only, no travel times
A travel time includes driving time and everything else during the journey. The TD matrix does not include maneuver time near site, breaks, refueling, site search, etc. This means that if customers want extra driving time between two stops, they will need to add some extra time above our driving times.