[pgrouting-users] Wrong results at routing by choosing wrong start/end points?

[Marco Lechner - FOSSGIS e.V.]

Hi,

regarding to http://pgrouting.postlbs.org/wiki/killing_some_bugs

by calculating the Route twice by changing start and endpoint ist a 
possibility to fix the problem in the triangle-like way at the 
screenshot. But does this also work if the points are many crossings 
away? Dont' I have to compare four results?:

SELECT gid, the_geom FROM dijkstra_sp_delta(
     '".TABLE."', ".$startEdge['source'].", ".$endEdge['target'].", 3000)

SELECT gid, the_geom FROM dijkstra_sp_delta(
     '".TABLE."', ".$endEdge['source'].", ".$startEdge['target'].", 3000)

SELECT gid, the_geom FROM dijkstra_sp_delta(
     '".TABLE."', ".$startEdge['source'].", ".$startEdge['target'].", 3000)

SELECT gid, the_geom FROM dijkstra_sp_delta(
     '".TABLE."', ".$endEdge['source'].", ".$endEdge['target'].", 3000)

Marco

P.S. for own reference: 
http://pgrouting.postlbs.org/wiki/WorkshopOL2.7andOSM


Kai Behncke schrieb:
> Hi Marco,
> 
> [...]
> If I don`t misundestand your question the "killing_some_"bugs"-part of the tutorial 
> http://pgrouting.postlbs.org/wiki/killing_some_bugs (and further documents)
> [...]
> If not please write :-)
> 
> Best regards, Kai
> -------- Original-Nachricht --------
>> Datum: Fri, 23 Oct 2009 11:24:31 +0200
>> Von: "Marco Lechner - FOSSGIS e.V." <marco.lechner at fossgis.de>
>> An: pgrouting-users at lists.postlbs.org
>> Betreff: Wrong results at routing by choosing wrong start/end points?
> 
>> Hallo Kai, hi list,
>>
>> because I'm trying to enhance our pgrouting-based Routing-application I
>> took a look at these webpages, because they fit our needs (1. clicking
>> somewhere on the map to set start and endpoint, 2. finding the nearest
>> points on the routing network, 3. calculating the shortest path from
>> start_on_road to end_on_road:
>> http://pgrouting.postlbs.org/wiki/LoadingtheCode3 and
>> http://pgrouting.postlbs.org/wiki/LoadingtheCode4
>>
>> Using the code on "LoadingtheCode4" to fill the gaps between the routing
>> result on "LoadingtheCode3" is good. But one question appeared:
>> are the start- and endpoints calculated on "LoadingtheCode3" really on
>> the shortest path from my_start_point to my_end_point?
>> The routing is calculated by using the nearest edge from the clicked
>> my_start|end_point. But because the real start- and endpoints are a
>> distance away from the start|end_on_road this could modifiy the result
>> in choosing a short , but not the shortest way.
>>
>> See attached sketch for explanation. Am I getting anything wrong or is
>> this really an "error"?
>>
>> Marco
>>
>> -- 
>> +++++++++++++++++++++++++++++++++
>>
>> FOSSGIS e.V.
>> die unabhängige Hilfe bei freier GIS-Software und freien Geodaten
>> www.fossgis.de
>>
>> +++++++++++++++++++++++++++++++++
>