Page 1 of 1

Brenner-pass mistery

Posted: Wed Sep 30, 2020 12:05 pm
by innomedio
Dear Bernd,

We've found something interesting, regarding the violations on the A13 Motorway, from the Brenner-pass.

For both routing we've used the same, default mg40t profile.

Case 1 - pic1
- from: Bolzano
- to: Innbruck
- without turning on the allowance of the violations the algorithm won't let us go trough the Brenner-pass

Case 2 - pic1
- from: Bolzano
- to: Innbruck
- if we allow the violations, then it let us go trough the Brenner-pass, but says that we have a violated segment: AT 6156 Gries am Brenner Brenner-Autobahn, 7,5T weight violation

Case 3 - pic2
- from: Vipiteno
- to: Innbruck
- however, if we route from Vipiteno, than the route goes through the Brenner-pass, with the same profile, without any violations

Pic1:
https://innomanagement.hu/ptv/2020-09-30/pic1.png

Pic2:
https://innomanagement.hu/ptv/2020-09-30/pic2.png

How is it possible?

We have one tip for this, but we're not sure:
- if our information is correct, the A13 motorway in AT is closed for >7.5 truck between 22:00-5:00 hours
- but the whole route is about 2 hours long, and we've calculated at 13:00

Please help us!

Thank you,
Peter

Re: Brenner-pass mistery

Posted: Wed Sep 30, 2020 12:50 pm
by Bernd Welter
Hello Peter,

I have to create a ticket for this: please send the complete requests to bernd.welter@ptvgroup.com

Best regards,
Bernd

Re: Brenner-pass mistery

Posted: Fri Oct 02, 2020 11:45 am
by Bernd Welter
Hello Peter,

I tried to reproduce the requests in the cloud:
https://xroute-eu-n-test.cloud.ptvgroup ... /ws/XRoute

Looks like the server performs a very little detour with the third request
Left: route #2<br />Right: Route #3
Left: route #2
Right: Route #3
Does this explain the behaviour? The sub routes are not really equal ;-)

Best regards,
Bernd

PS: images are animated, click on them)

Re: Brenner-pass mistery

Posted: Fri Oct 02, 2020 1:33 pm
by Bernd Welter
Now that was a big challenge.... First I didn't get what exactly you wanted to have but I think I get the point now...

As I mentioned in the previous post under some circumstnces the router takes a local detour to avoid a violation.
Detour through a NC_5 segment
Detour through a NC_5 segment
The detour requires the engine to use a segment of network class NC_5. As this is a minor category the scope (levelling) is a relevant factor here:
The regular search space considers NC_5 with a scope of "20" which means "do not consider NC_5 segments if they are further away from any waypoint that "20" (kilometers).

So when at least one waypoyint is close enough (scenario #3) the detour is "possible" because the NC_5 segment is "known".

If you start your route in Bolzano (too far away from the relevant NC_5 segment) with the scope of "20" (or even "30") the segment is "invisible" and the engine neeeds to perform the detour to avoid the conflict/violation. By setting the Scope to about "35" the scope is large enough.
scope set to 35: the detour is &quot;visible&quot;
scope set to 35: the detour is "visible"
Ok with that?

Best regards,
Bernd

Re: Brenner-pass mistery

Posted: Mon Oct 05, 2020 4:43 am
by innomedio
Dear Bernd,

Thank you for your help! Now it all makes sense. :)

We're quite new to the xServer and there are so many parameters considered in routing.

At this example, we've used the default mg40t profile. The profile includes these levelling rules.

We didn't know that, if you want to cross the Brenner-pass in that direction, than you have to use a low level road.

We'll implement these levelling values, as parameters to our GUI, so that the users can change them during routing.

Thanks again!
Peter

Re: Brenner-pass mistery

Posted: Mon Oct 05, 2020 8:33 am
by Bernd Welter
feel free to check this article about LEVELLING, too ;-)