[pgrouting-dev] Ask how to use : Euclidean distance heuristic in A* in pgrouting

[Sanak]

Hi Robi,

I am not familiar with A* heuristic logic,
but if you want to build pgRouting Windows binary,
see following source and docs.

https://github.com/sanak/pgrouting
https://github.com/sanak/pgrouting/blob/mingw/BUILD.mingw
https://github.com/sanak/pgrouting/blob/mingw/BUILD.mingw64

And if you want to skip building dependency libraries,
use following build environment.

http://sdrv.ms/NFCgr2
 +- Build.zip     ... for 32bit work folder. (extract to "C:\")
 +- MinGW.zip ... for 32bit build environment. (extract to "C:\" and run
"C:\MinGW\msys\1.0\msys.bat")
 +- Build64.zip ... for 64bit work folder. (extract to "C:\")
 +- ming64.zip ... for 64bit build environment. (based on PostGIS one.
extract to "C:\" and run "C:\ming64\msys\msys.bat")

Regards,

2012/7/10 Stephen Woodbridge <woodbri at swoodbridge.com>

> On 7/9/2012 9:24 PM, Haruhi Suzumiya wrote:
>
>> Good morning,
>>
>> Dear mr/mrs, :)
>>
>> I'm robi, my username is haruhi. I'm newbie in pgrouting.
>> I want to try use euclidean distance heuristic in A* in pgrouting, may u
>> give me tutorial how to change the heuristic and recompile it? :)
>> Fyi, I'm use windows 7, postgre 1.8.4, pgRouting-1.02_pg-8.3.3 :)
>>
>
> Hi Robi,
>
> The pgRouting package already supports A* and by definition it uses a
> euclidean distance heuristic. See the function shortest_path_astar(), there
> also may be some wrapper function with names like astar_sp_*
>
> The wrapper function are just higher level convenience function.
>
> There are a couple of tutorials around that probably have good examples of
> how to use the functions.
>
> If you want to change the heuristic, then you will need to look at the
> Boost Graph code that is used to implement it. We don't have any
> documentation on the implementation, but there are a lot of docs on Boost
> Graph and you can get help from that list.
>
> What I would do is write a wrapper for the Boost Graph code and reads a
> text file to input the graph, sets up the structures to pass the data to
> algorithm to run it and then prints out the results. You can do all this
> outside of the database which greatly simplifies debugging your changes.
> When you have it working the way you want, then it should be a clean drop
> in replacement for the existing code or you can clone the existing code and
> rename it for the new algorithm.
>
> -Steve W
> ______________________________**_________________
> pgrouting-dev mailing list
> pgrouting-dev at lists.osgeo.org
> http://lists.osgeo.org/**mailman/listinfo/pgrouting-dev<http://lists.osgeo.org/mailman/listinfo/pgrouting-dev>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pgrouting-dev/attachments/20120710/72b00301/attachment.html>