Simulations · SUMO 1.22.0

Munich centre, seed 408

Synthetic random-trip demand, chosen so that the reference run clears; only 07:00 to 08:00 departures are scaled.

06:30Loading the recorded runs…

Zoom with the mouse wheel or the buttons, drag to move, double-click to see the whole city again. Both panels move together.

Reference demand · scale 1.000 · seed 408

3170 / 3170 arrived by 09:00.

vehicles on the network (all)
sampled cars drawn
mean speed of sample, km/h
sampled cars below 2 km/h

Llama-scaled demand · scale 1.098 · seed 408

3446 / 3502 arrived by 09:00, 55 still on the network, 1 never inserted.

vehicles on the network (all)
sampled cars drawn
mean speed of sample, km/h
sampled cars below 2 km/h

moving, above 30 km/hslow, 2 to 30 km/hstopped, below 2 km/hbus

All demand variants

Demand variantscaleCohort cars per seedTime loss, completed trips (s)Elapsed, all scheduled cars (s)Unfinished, all seedsCompletion, minimum
Reference demand1.0003170-3170241.8591.00100.0%
Constant-only fallback0.9703071-3083241.6591.40100.0%
Gemma-scaled demand1.0283253-3265241.5590.70100.0%
Llama-scaled demand1.0983452-3502246.2621.55698.4%
GPT-OSS-scaled demand1.0103198-3204241.7590.70100.0%

Means over the seeds listed per variant. Time loss counts completed cohort trips only; elapsed time counts every scheduled car to its arrival or to the 09:00 horizon, so it does not hide unfinished trips.

Where these runs come from

Road network from OpenStreetMap (ODbL 1.0, via Overpass), netconvert; SUMO 1.22.0

15% of vehicles report their position every 30 s (SUMO floating-car output, geographic coordinates). time-to-teleport -1, ignore-junction-blocker 20, step 1 s, 06:30 to 09:00.

These are sensitivity runs on technical scenarios, not calibrated city forecasts. Every number on these pages comes from the recorded SUMO output; the site adds nothing.

Download the recorded positions (JSON)