Distributed Evolutionary Computation using REST
Upcoming SlideShare
Loading in...5
×
 

Distributed Evolutionary Computation using REST

on

  • 635 views

Presentation for IWDECIE within CEC2011

Presentation for IWDECIE within CEC2011

Statistics

Views

Total Views
635
Views on SlideShare
635
Embed Views
0

Actions

Likes
1
Downloads
4
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as OpenOffice

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment
  • Picture taken from http://www.flickr.com/photos/joeschram/207822199/in/photostream/
  • Picture from http://www.flickr.com/photos/25653307@N03/2716213900/in/photostream/
  • Not all clients all the same, so no linear speedup should be expected. In any rate, the server is bound to have a threshold over which it should overload itself
  • Picture from http://www.flickr.com/photos/beauteous/48847025/in/photostream/
  • All source, data sets, experiment results for this paper are available from Sourceforge (in fact, they were while we were doing it). Source is also available from the CPAN Perl module server worldwide, in two separate modules: the algorithm itself it's available from the author's web site

Distributed Evolutionary Computation using REST Distributed Evolutionary Computation using REST Presentation Transcript

  • Distributed Evolutionary Computation using REST P.A. Castillo, M.G. Arenas, A.M. Mora, J.L.J. Laredo, G. Romero,V.M Rivas J.J. Merelo GeNeura Group http://geneura.wordpress.com http://twitter.com/geneura
  • If you've been paying attention REST beats other web-services implementations: lighter, more scalable
  • Let's put it to work Using a single server, different clients (master/slave implementation)
  • What problem?
    • Optimization of topology + initial weights for a multilayer perceptron: G-Prop algorithm (Castillo et al., 1999)
    • Evaluation lengthy and time-consuming
  • Client/server implementation
  • Old McREST had a farm
    • But we will not be using it for this.
    • Populations running in parallel on clients and use REST server for communication.
    • Hope to achieve better scaling that way.
  • Reminder: REST architecture Send chromosomes to the server Request chromosomes Send chromosomes Get evaluated chromosomes Uses PerlDancer Uses Algorithm::Evolutionary
  • Evolutionary algorithm parameters
  • Experimental results
  • Speedup nodes speedup
  • REST your parallelization in peace Just add clients!
  • Open source your science!
  • Any question? http://geneura.wordpress.com http://twitter.com/geneura