[Soc] TDSP implementation for pgRouting Weekly Report

[Jay Mahadeokar]

Hi,

Weekly Report #11
Date - 12th Aug 2011

Current Progress.
---------------------------------------------------------------------------------------------------------------------------------------------------------
I completed the APSP algorithm and removed a few bugs and problems that were
present in the previous draft.
Updated the github repository, I also wrote a tutorial to get started, run
and test the algorithm. [2].

For TDSP algorithm, I polished the code, added comments etc. Also, I wrote a
plsql data generator function that generates random time dependent data
corresponding to the ways table. It demonstrates the tdsp algorithm in a
better way. I have added the plsql function to github repo too. [3].

Next weeks goal
---------------------------------------------------------------------------------------------------------------------------------------------------------
Write tutorial for testing using random data, work on documentation etc.
Keep doing some tests. Discuss about packaging the algorithm with pgRouting
library.

[1]https://github.com/pgRouting/pgrouting/tree/gsoc-tdsp
[2]https://github.com/pgRouting/pgrouting/wiki/APSP
[3]
https://github.com/pgRouting/pgrouting/blob/gsoc-tdsp/extra/tdsp/sql/data_generate_tdsp.sql
On Fri, Aug 5, 2011 at 10:58 AM, Jay Mahadeokar <jai.mahadeokar at gmail.com>wrote:

> Hi,
>
> Weekly Report #10
> Date - 5th Aug 2011
>
> Current Progress.
>
> -------------------------------------------------------------------------------------------------------------------------------------------------------
> After much brain-storming in the community, we decided that the proposed
> cyclic data optimization was a little bit premature, and we would just focus
> on completion of the basic algorithm as proposed in the GSoc proposal during
> the GSoc timeline.
>
> So, the major function will be pgr_time_dependent_shortest_path.
> For cyclic data, we will have function like: pgr_tdsp_cyclic. The
> weight_map class has been modified for that, but the default constructor
> keeps the behaviour non-cyclic for now. We will keep brainstorming until the
> requirements and idea of cyclic data are properly identified and defined and
> then provide support for the optimization.
>
> In the mean time I also cleaned up the code and continued testing.
>
> Next week goals.
>
> ------------------------------------------------------------------------------------------------------------------------------------------------------
> Try and finish up the documentation work needed for tdsp. Including
> examples to convert some common data formats to tdsp and potential plsql
> scripts for the same.
>
> Since the major portion of GSoc proposal is completed,I also plan to work
> on all-pairs shortest path implementation for pgRouting(not part of the GSoc
> proposal, I had worked on it earlier), and I will try and finish it up and
> document it within GSoc timeline as planned earlier with my mentor if time
> permits.
>
>
>
>
>
> --
> Regards,
> -Jay Mahadeokar
>
>


-- 
Regards,
-Jay Mahadeokar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/soc/attachments/20110812/80feb067/attachment.html