Re: [sumo-user] How would Time Step Length influence simulation?

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

Hello,

The main differences of lower step length are:

1. smoother trajectories

2. increased running time (since more steps must be computed)

3. different insertion pattern since insertion only happens only at step end

Traffic lights near congestion can be quite chaotic though. The first thing to check is whether the difference in behaviors holds up if you repeat the simulation with different seeds.

You can easily test this with

sumo/tools/runSeeds.py -a sumo -k case1.sumocfg,case2.sumocfg --seeds 0:100 --statistic-output stats.xml

then run

sumo/tools/output/attributeStats.py case1.sumocfg_0/*.stats.xml

sumo/tools/output/attributeStats.py case2.sumocfg_0/*.stats.xml

Am Mo., 13. März 2023 um 22:45 Uhr schrieb George Li <wzhli2020@xxxxxxxxxxx>:

Hi Community,

I am writing this email to ask about how would the step-length value in the sumocfg file would influence the simulation result. I tried two values as below:

Case 1:

<time>

        <step-length value="0.1"/>

 </time>

Case 2:

<time>

        <step-length value="1"/>

 </time>

All other settings are the same. I run simulations using a fixed-time traffic signal controller at an isolated intersection. However, the results under the two cases are significantly different: the average total travel loss of vehicles for case 1 is around 40s but for case 2 is around 80s. From GUI, congestion can be observed under Case 2 but not for Case 1. May I ask you about possible reasons for that? Thank you.

Have a nice day.

Best,

Wangzhi

_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user