Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce, mail # user - Ant Colony Optimization for Travelling Salesman Problem in Hadoop


Copy link to this message
-
Re: Ant Colony Optimization for Travelling Salesman Problem in Hadoop
Steve Lewis 2012-05-05, 15:20
yes - if you know how you can put it in distributed cache or if it is small
put in as a String in the config or have all InputFormats read it from
somewhere

On Sat, May 5, 2012 at 8:08 AM, sharat attupurath <[EMAIL PROTECTED]>wrote:

>  I looked at both the files. in AbstractNShotInputFormat it is mentioned
> that this input format does not read from files. My input is in a text
> file. I want the whole file as a single record. So is it enough if i just
> copy the contents of the file and return it as a string from
> getValueFromIndex() ?
>
> ------------------------------
> Date: Fri, 4 May 2012 13:15:46 -0700
>
> Subject: Re: Ant Colony Optimization for Travelling Salesman Problem in
> Hadoop
> From: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED]
>
> Look at NShotInputFormat
> call setNumberKeys to set the number of ants then
> change the method getValueFromIndex to return the text representing the
> original problem to each mapper -
> what happens in the second and third jobs is an exercise but the saved
> test needs to have pheremones and the original problem
>
>
> On Fri, May 4, 2012 at 9:54 AM, sharat attupurath <[EMAIL PROTECTED]>wrote:
>
>  Hi,
>
> Thanks a lot for the quick reply sir! We are new to Apache Hadoop and
> haven't yet understood it properly yet. Can you please elaborate on how we
> can have multiple stages of mapreduce jobs for combining the trails as you
> have mentioned?
>
> We have been trying to find out how to write a custom splitter and almost
> all online resources tell to subclass the FileInputFormat and write only a
> custom RecordReader? Will it be possible to generate our splits in that way?
>
> Regards
>
> Sharat
>
> ------------------------------
> Date: Fri, 4 May 2012 09:22:51 -0700
> Subject: Re: Ant Colony Optimization for Travelling Salesman Problem in
> Hadoop
> From: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED]
>
>
>
>
> On Fri, May 4, 2012 at 9:06 AM, sharat attupurath <[EMAIL PROTECTED]>wrote:
>
>  Hi,
>
> We are trying to parallelize the ant colony optimization algorithm for TSP
> over hadoop and are facing some issues. We are using TSPLIB as input files.
> The input is a text file containing eucledian coordinates of the cities -
> first column is city number and the next two columns contain the x and y
> coordinates respectively.
>
> What we intend to do is to take input from this single file, send copies
> of it to multiple mappers (each mapper acts like the ant in the algorithm),
> each mapper works on its input to find its own TSP solution that it outputs
> and finally the reducer outputs the smallest tour found by the mappers.
> Hope we are in the right track. Here are the issues:
>
> 1) Since the input file is small, we need to force hadoop to fire up
> multiple map tasks by replicating the input. How can we make an InputSplit
> of the whole file and replicate it so that the input can be sent to
> multiple mappers?
>
> Write a custom splitter sending the same data to all mappers - the only
> critical criteria is the number of "Ants"
>
>
> 2) the algorithm uses a shared pheromone array and each mapper needs to
> read and write data from this. How can we share the pheromone data across
> the mappers.
>
> You cannot share data across mappers and should not attempt to do so -
> Better to use the reducer(s) to combine trails  combine first pass trails
> and
> then pass the combined trails to another mapreduce job - this with the
> original problem plus the current pheremone trails
>
>
> Hope the questions are clear enough. Any help would be greatly
> appreciated.
>
> Thank you
>
> Regards
>
> Sharat
>
>
>
>
> --
> Steven M. Lewis PhD
> 4221 105th Ave NE
> Kirkland, WA 98033
> 206-384-1340 (cell)
> Skype lordjoe_com
>
>
>
>
>
> --
> Steven M. Lewis PhD
> 4221 105th Ave NE
> Kirkland, WA 98033
> 206-384-1340 (cell)
> Skype lordjoe_com
>
>
>
--
Steven M. Lewis PhD
4221 105th Ave NE
Kirkland, WA 98033
206-384-1340 (cell)
Skype lordjoe_com