[Routing] Routing algorithms

[Marcus Wolschon]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Nic Roets schrieb:
> (So why didn't Stefan say that in the first place ...)
>
> The turn restriction problem is easy to solve once you realize that
>  OSM nodes need not correspond to graph vertexes. The graph
> vertexes are states (search google for state machine).
>
> Let's say node N1 is part of  one simple turn restriction relation
> W1,N1,W2 and one complicated turn restriction relation
> W3,N2,W4,N1,W5. Then we would associate 3 states with N1 : N1a : N1
> was reached after traveling along W1. N1b : N1 was reached after
> traveling along W3, then N2, then W4. N1c : N1 was reached in some
> other way.
>
> And N2 will have at least 2 states associated with it.


Yes, that will work for me. Thanks for the idea!
Expand a node into a small graph with one-way
- -edges that represents the same rule.

This however is my problem. Stefan has a different
issue. His metric depends on the metric of the path
used to reach the currently evaluated node.

Marcus
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFI1p33f1hPnk3Z0cQRAtfTAJ4p73mDHfTnrfWkwI5cPskf78FVnwCeMYzd
WLWLzSTZAZ15NfBwfvg7qHI=
=B560
-----END PGP SIGNATURE-----