[Graphhopper] start and final routing point
Peter K
peathal at yahoo.de
Tue Mar 12 13:18:10 UTC 2013
Hi,
this is the limitation of Location2IDQuadtree which is mentioned in this
issue:
https://github.com/graphhopper/graphhopper/issues/17
The good news is that I'm working on this because the current
alternative Location2IDIndex implementations are too memory intensive.
As a workaround you could try to increase the size used for prepareIndex
(see e.g. GraphHopper.initIndex)
> Graphhopper runs in acceptable speed and needs very less memory that
OSRM for example.
Do you have any numbers for comparison? OSRM also supports turn
restrictions (see issue #2 for graphhopper) where the differences could
come from.
Regards,
Peter.
> Hello,
>
> it seems, that graphhopper is not optimal in retrieving the best start
> and endpoint on the most suitable street. At the moment, start und
> endpoint are only start or end point on osm ways.
>
> At short dinstances the result is vary. There can be several douzens
> or 100 meters between the real world routing and the calculated one.
>
> Graphhopper runs in acceptable speed and needs very less memory that
> OSRM for example. So i want to go on further for my mass queries.
>
> Is there a solution for this or a parameter?
>
>
> _______________________________________________
> GraphHopper mailing list
> GraphHopper at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/graphhopper
--
GraphHopper.com Your way is our destination!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/graphhopper/attachments/20130312/f4fe761d/attachment.html>
More information about the GraphHopper
mailing list