The default speedFactor distribution is a little aggressive

[woodwm]

Add your issue description here.

In the documentation, here

The default for passenger cars is "normc(1, 0.1, 0.2, 2)" which implies that ~95% of the vehicles drive between 80% and 120% of the legal speed limit.

I think the lowerCutOff and upperCutOff for the default value are too aggressive. This may cause a vehicle running very slow on the road and block the following vehicles in a signle lane simulation, for example #13956. The distribution of speedFactor is actually the distribution of maximum speed rather than the speed. I think too large curoff range is a problem in the simulation.

I understand changing the default value is a big decision. Please consider modifying them when possible, for example changing to "normc(1,0.1,0.8,1.2)".

If possible, upload an example that shows your problem.

SUMO-version: 1.18.0 operating system:

[m-kro]

I want to recall how the standard normal distribution density looks like. If our speed factor implementation is done right, it should select such low speed for at most 2% of the vehicles.

[woodwm]

@m-kro In the default truncated normal distribution, 2.28% vehicles get the speedFactor lower than 0.8. It looks small, but this means 57 vehicles get the speedFactor lower than 0.8 in a simulation of 1h given the capacity of 2500 veh/h. Every such vehicle may create a slow moving jam. I do not think such low maximum speed helps in the simulation.

0.2 as the lower cutoff is too small in my mind.

[m-kro]

There are other scenarios (motorway without speed limit / with rather high speed limit) where this speed factor distribution is fine. I don't think we will get to an optimal solution here.

[behrisch]

maybe we should find a source or check the data Peter has already analyzed