Developer.Matrix Routing : Traffic Modes comparison

[Bernd Welter]

Hi there,

Technical Concept "TrafficMode"

as the handling of

• TrafficMode: NULL / AVERAGE / REALISTIC
• StartTime: NULL / hh:mm...
• SYNC / ASYNC

seems to be confusing sometimes I decided to prepare some fancy example.
Take your time to look into the charts to learn from the overview...

Coordinates are been based on
48.9236907958984,8.47517967224121
48.9424781799316,8.40237045288086
49.0109596252441,8.40847969055176

• TrafficMode = null , Start Time = null
• TrafficMode = AVERAGE , Start Time = null
• TrafficMode = AVERAGE , Start Time = 16:00
• TrafficMode = AVERAGE , Start Time = 23:00
• TrafficMode = REALISTIC , Start Time = 16:00
• TrafficMode = REALISTIC , Start Time = 23:00

• TrafficMode=NULL produces equal results as TrafficMode=Average (if StartTime == null)
• TrafficMode=Average, StartTime=NULL produces more relaxed results than TrafficMode=Average, StartTime=16:00
• TrafficMode=Average, StartTime=23:00 produces more relaxed results than TrafficMode=Average, StartTime=16:00
• TrafficMode=Average, StartTime=23:00 produces more relaxed results than TrafficMode=Average, StartTime=NULL
• Average(16:00) produces more relaxed results than Realistic(16:00)

• Start Time given or not has a clear impact on Calculation Time

Whenever a start time is given the calculation time is way higher than without a start time

Best regards,
Bernd

[Bernd Welter]

I made some more tests...
Combined

• various (square) matrix sizes
• complexity
  • average without start time
  • average with start time
  • average with toll time
  • realistics with start time
• protocols: sync versus async

Please be aware that not every combination is supported (as of today, 21.3.2025).
Here are some impossible ones:

• SYNC + AVERAGE + StartTime (no matter what size)
• ASYNC + REALISTIC + TollTime (no matter what size)
• SYNC + AVERAGE + extended size (>250'000 relations)

History of my calls

Calculation times somewhere clearly below 60 seconds