Re: [sumo-user] Realistic Simulation with traffic data and calibration

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

Hello,

- generating more candidate routes as input for routeSampler is a viable approach. The only downside is that the optimization becomes slower as you add more routes

- manually adding a route that passes a specific sequence of edges is fine as well (though you would often want more than a single route for each counting location)

- the integer argument to option --optimize restricts how much the usage count for each route may be changed. This is to prevent the case where a single route is used all the time to the detriment of other routes.  The idea behind this is that the input routes carry some information regarding the desired distribution of routes and the optimization must find a compromise between satisfying the intended distribution as well as the counts.

regards,

Jakob

Am Mi., 1. Apr. 2020 um 06:23 Uhr schrieb Tetris <schmelter@xxxxxxxxxxxxxxxxxxx>:

Hello,

I used the method with non-consecutive turning relations.

I got the warning "no routes pass edge" again for a few edgeRelations. But
as I created more random routes with randomtrips.py and p 0.7 instead of p
1, there was no warning. Is this actually a problematic approach?  I
couldn't solve the warning otherwise. Or should I create the missing route
myself and put it in the random routes file?

The result with p 0.7 was GEH<5 for 91 %.

Wrote 9545 routes (2762 distinct) achieving total count 16695 at 83
locations. GEH<5 for 91.57%
Warning: underflow locations: count 43, min 1.00 (('24654525#8',
'28497111#0')), max 75.00 (('-4540284#2', '4540284#0')), mean 13.98 Q1 2.00,
median 8.00, Q3 21.00 (total 601)
Warning: overflow locations: count 27, min -93.00 (('25184203#0',
'-24678549#1')), max -1.00 (('182604676#2', '182603253#2')), mean -14.30 Q1
-16.00, median -5.00, Q3 -2.00 (total -386)

I wanted to improve the result by using --optimize-input  on the output.
(https://sumo.dlr.de/docs/Tools/Turns.html#further_calibration)

routeSampler.py -r  outputOfRS.rou.xml -t turncounts_passenger.xml
--turn-max-gap 8 --optimize-input  -o outputOfRS_cal.rou.xml

But:
Option --optimize-input requires an integer argument for --optimize

What does the integer argument stand for? How should the command look like?
I used "--optimize full" before and got
"OptimizeWarning: x0 is used only when method is 'revised simplex'.
  warn(warning_message, OptimizeWarning)
Optimization succeeded".




Jakob Erdmann wrote
> 
> 3) In the latest development version, routeSampler provides for two
> different methods to define non-consecutive turning relations (
> https://sumo.dlr.de/docs/Tools/Turns.html#generalized_route_restrictions)





--
Sent from: http://sumo-user-mailing-list.90755.n8.nabble.com/
_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user