Re: [sumo-user] Meso Sumo TauJF

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

Hello,

the tauff values use in mesosim denotes net-time headways (front-bumper to rear-bumper).

The gross headways (front-bumper to front-bumper) will be computed from tauff by taking into account road speed and vehicle length.

This means you do not need to make any adjustments for road speed. (The original Eissfeldt model used tauff for gross headways and thus had problems with networks with heterogeneous speeds).

The idea of https://github.com/eclipse/sumo/issues/5745 is to allow localized calibration of the model to reflect anomalous situations but this is not yet implemented.

The taujf value, again denotes the net-time headway when leaving a jam. In my opinion this is something to be measured from real-world data rather than computed. It is a free model parameter after all.

regards,

Jakob

Am Di., 3. Dez. 2019 um 14:02 Uhr schrieb Flitsch Christina <Christina.Flitsch@xxxxxxxxxxx>:

Dear SUMO-Team,

I’m searching for two inputs about the Tau-Meso settings.

 

In general my network is quite large – now I’m using a small testnetwork for finding well-fitting parameters, I’ve tried to add every speed type etc. I need into this testnetwork.

 

I’ve different speeds (there are urban, suburban, rural areas and highways) and a lot of different edge lengths.

è Are there any suggestions for the calculation of tau which Vmax I should use (the total Vmax of the hole network – the Median, Q75 etc.?), the smaller the Vmax, the higher the tau-ff

è Or is there a way that this depends on the lane type itself? (I didn’t really know how to handle this information https://github.com/eclipse/sumo/issues/5745)

 

The second question is, that I’m not totally sure how to calculate the tau jam-free and I would like feedback on the tau_ff if possible:

è For the tau_ff: https://github.com/eclipse/sumo/issues/2244 and https://www.eclipse.org/lists/sumo-user/msg04886.html  were looked at. I’ve tried the tau_ff once with (lcar+lgap)/speedmax and once with the formula max. flow=1/(tau_ff+(lcar+lgap)/speedmax) and I receive the same results - but definite a lot lower than the tau-values Eissfeld uses in his examples (ex. I calculated 25m/s tau 0.3 (Eissfeld has an example with 19.6m/s with tau_ff 1.4s, for 16m/s I calculated a tau with 0.46 and for 9.7m/s I calculated a tau with 0.77)– do you know if this results are correct, though they are much lower?

è Can you give me a hint how to calculate the jam-free? (I’ve looked at the thesis of Eissfeld, I’ve found that once there he sets the value) However, I’m unsure how to calculate it when I do not have a jam-free or jam-jam value

 

Also I wanted to note that in the thread https://www.eclipse.org/lists/sumo-user/msg04884.html the following is pointed out: “The other point is that in the dissertation on page 70 formula 3.24 there is a relationship between tau_jf and tau_jj defined indicating that tau_jj is bigger than tau_jf but in SUMO it is the other way around” this is not correct – Eissfeld writes “tau_ff<=taujf” which is handled exactly that way in the default values and is important for a static jam.

 

Thanks in advice!

Greetings,

Christina

_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/sumo-user