Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [sumo-user] Error using load-state option with calibrators
Hello,

1) options are fine. I would not expect a 90mb simulation network to drop below real-time simulation speed. Possibly your simulation spends a lot of time computing routes. Check if either of the following options speeds things up:
--routing-algorithm astar
--routing-algorithm CHWrapper
--device.rerouting.threads <INT>

2) Car following models are not used in mesosim. Slope is taken into account in mesosim only for emission/fuel consumption.

3) The error indicates that a vehicle with the same ID already exists in the simulation. This may be a bug but I would need a (small!) failing example to look into this.

regards,
Jakob

Am Mo., 25. Nov. 2019 um 12:29 Uhr schrieb Florian.Berkes@xxxxxxxxxx <Florian.Berkes@xxxxxxxxxx>:

Hi Jakob,

 

thanks for your quick reply and fixes.

 

“Regarding vehicles with reroute="0". Can you please provide a minimal sample input that reproduces this issue?”

Sharing the network file will be difficult. The file is about 90 MB. Any suggestion how we can provide a smaller subset?

 

The meso simulation takes also a quite long time, we thought it will be much faster.

Currently, 1 hour simulation takes about 24h real time.

 

  • We use the options --mesosim true --meso-junction-control.limited true --meso-overtaking

Are they fine?

 

  • We also use the carfollowing model KraussPS to consider slopes. The altitude was extracted from SRTM data.

Is this model used if we simulate with mesosim?

 

  • We still have an error with an calibrator after we load the state file from the previous:

Loading state from 'state_25200.00_fixed.xml'... done (18747ms).
Loading done.
Simulation version 1.3.1 started with time: 25200.00
Error: Emission of vehicle 'cali_133849585.25233991.0' in calibrator 'cali_133849585'failed!
Quitting (on error).
DijkstraRouter answered 1014 queries and explored 43169.76 edges on average.
DijkstraRouter spent 42128ms answering queries (41.55ms on average).
Segmentation fault

 

Thanks again for your help,

Florian

 

 

--

Dr. Florian Berkes

umlaut

cell +49 151 27654 587

florian.berkes@xxxxxxxxxx

 

www.umlaut.com

 

umlaut solutions GmbH

legal disclaimer, mandatory information and the umlaut office addresses can be found at

imprint.umlaut.com/SOL

Information about the processing of your personal data can be found here.

 

Von: sumo-user-bounces@xxxxxxxxxxx <sumo-user-bounces@xxxxxxxxxxx> Im Auftrag von Jakob Erdmann
Gesendet: Samstag, 23. November 2019 09:19
An: Sumo project User discussions <sumo-user@xxxxxxxxxxx>
Betreff: Re: [sumo-user] Error using load-state option with calibrators

 

Hello,

the 'depart' value within the statefile encodes departDelay. This can become negative due to https://github.com/eclipse/sumo/issues/964

The error you are seeing comes from our xml-schema validation (https://sumo.dlr.de/docs/XMLValidation.html)

I've updated the schema files to allow negative departDelay.

To prevent the error you can either

- disable xml validation (--xml-validation never)

 

Regarding invalid routeDistribution: Thanks for letting me know. This is now fixed: https://github.com/eclipse/sumo/issues/6308

Regarding vehicles with reroute="0". Can you please provide a minimal sample input that reproduces this issue?

Regarding lane placement: Mesosim mostly does not take lanes into account (except turn lanes at intersection). Placement of vehicles in the rightmost lane is therefore purely a visual issue.

 

regards,

Jakob

 

Am Fr., 22. Nov. 2019 um 09:22 Uhr schrieb Florian.Berkes@xxxxxxxxxx <Florian.Berkes@xxxxxxxxxx>:

Hi Jakob,

 

we run a 24h scenario’s using sumo meso (v1.3) using

  • randomtrips generate for each hour,
  • multiple calibrators (> 1000) in 1min-frequency, only streets with prio 7 to 13 and larger than 100 m
  • using the option load-save state for every hour suffix (xml)

 

We had some issues with the calibrators attributes and did some fixes:

  • vehicle inserted by calibrators with attribute reroute=0. When loading the state, insertion of these vehicles fails. Fix: set reroute=1
  • some routeDistribution entries have attribute probabilities="", which results in an error when loading the state. Fix: removing these distributions

 

After 5am of the simulation, we faced the following error loading the state file:

 

Error: value '-4327.00' must be greater than or equal to minInclusive facet value '0'

​<delaybegin="959343"depart="-4327.00"end="932642"number="26119"time="125970483.00"/>

 

We deleted the line in the state file and continue the simulation.

However at 7 am we faced again an error for one calibrator.

 

In the GUI we see that most of the vehicles are mostly placed at the right lane (effect of meso simulation?)

Any idea what could be wrong with the calibrators?

 

Thanks

Florian

 

_______________________________________________
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

_______________________________________________
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

Back to the top