Difference between revisions of "Plan Marco A. Montes de Oca"

From IridiaWiki
Jump to navigationJump to search
 
Line 1: Line 1:
 
= Plan (future work) =
 
= Plan (future work) =
  +
* Still to define it clearly but something related with Particle Swarm Optimization (PSO)
 
  +
1. Implementation of some (at least the most influential) variants of PSO.
  +
This implementation should be modular and flexible enough to generate
  +
ad-hoc(according to problems characteristics) PSO versions and
  +
incorporate hybrid approaches easily. The language of choice will be
  +
C in order to reuse code developed by others.
  +
2. Comparison of all the different PSO variants and other EC techniques
  +
including ACOR on a variety of benchmark problems. Some of them
  +
are already coded and some others are proposed in the literature. This
  +
will contrast with previous and current practices in which only a small
  +
number of benchmark problems are used.
  +
3. Propose a PSO variant based on the F-Race technique. If the implementation
  +
is modular, it should be possible to consider various aspects
  +
such as neighborhood topologies as variables to be tuned and therefore,
  +
we could get ad-hoc PSO algorithms for specific applications.
   
 
= Goals =
 
= Goals =

Revision as of 16:20, 30 November 2005

Plan (future work)

1. Implementation of some (at least the most influential) variants of PSO. This implementation should be modular and flexible enough to generate ad-hoc(according to problems characteristics) PSO versions and incorporate hybrid approaches easily. The language of choice will be C in order to reuse code developed by others. 2. Comparison of all the different PSO variants and other EC techniques including ACOR on a variety of benchmark problems. Some of them are already coded and some others are proposed in the literature. This will contrast with previous and current practices in which only a small number of benchmark problems are used. 3. Propose a PSO variant based on the F-Race technique. If the implementation is modular, it should be possible to consider various aspects such as neighborhood topologies as variables to be tuned and therefore, we could get ad-hoc PSO algorithms for specific applications.

Goals

  • See above

Things to do

Description Start date Deadline Time required status
Particle Swarm Optimization Self-study Nov 1, 2005 Continuous process - In progress
Implementation of some variants of the PSO algorithm Nov 20, 2005 Continuous process -
Preparing material for the Optimization Group meeting Nov 29, 2005 Dec 1, 2005 ~1 day or less

Papers to write

Title Journal/Conference targeted Start date Submission deadline


Referee activities

International journals

Journal # papers paper received on review submitted on

International conferences

Conference # papers paper received on review to submit before