Difference between revisions of "Plan Prasanna Balaprakash"

From IridiaWiki
Jump to navigationJump to search
 
(11 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
= Plan (future work) =
 
= Plan (future work) =
  +
'''Hybrid Algorithms for Stochastic COP'''
   
  +
Milestone I : Jan-Mar 2006
  +
*Incremental Local Optimiztion
  +
*Finalizing experimentations on Local Search (For Journal)
  +
*Paper submission to ANTS 2006
  +
  +
Milestone II :Mar-May 2006
  +
*Ideas - Boosted Sampling, Consensus, Expectation & Consencus
  +
*ACO/F-Race plus Local search experimentations (For Journal)
  +
*Paper submission to Hybrid Metaheuristics
  +
  +
May - Jun 2006
  +
*Vacation
  +
  +
Milestone III :July-Oct 2006
  +
*Transferring ideas from Simulation Literature mainly from Ranking and Selection
  +
*Priliminary work on hybridising exact and local search techniques
  +
  +
Milestone IV :Nov-Dec 2006
  +
*Journal paper (EJOR)
  +
*Paper submission to GECCO 2007
  +
  +
Milestone V :Jan-April 2007
  +
  +
*Visiting Gutjahr's lab
  +
*Goal: Hybridisation, both from simulation and exact techniques prespective
  +
  +
Milestone VI :May-June 2007
  +
*Priliminary results - Conference paper, Advance results - Journal article (Results from the work done at Vienna)
  +
  +
Milestone VII :July-Oct 2007
  +
*Putting together the results of two hybridisation
  +
*Min 1 to Max 2 Journal papers
  +
  +
Milestone VIII :Nov 2007 - Mar 2008
  +
*PhD thesis
   
 
= Goals =
 
= Goals =
   
  +
  +
= Weekly Meeting=
  +
'''21-02-2006'''
  +
  +
1) Experimental results - Incremental Local search
  +
  +
Plots should be generated for the following three cases
  +
  +
a) x- axis - iterations, y-axis - distance between two solutions which are randomly generated.
  +
b) x- axis -iterations, y-axis distance between the global best solution to the greedy solution (Two graphs, one before the local search and one after the local search)
  +
c) x-axis -iterations, y -axis distance between the new global best and the old global best
  +
  +
The structure of the experimental setting and results
  +
  +
a)We will show reinit with time as the stopping criterion is not promising
  +
b)We will show reinit with same number of iteration doesn't help
  +
  +
c)Then we will show the experiments on distances
  +
d)Additionally, we will show the results when alpha=0
  +
  +
  +
2)Discussion - Structure of the paper ANTS-2006
  +
  +
3)Future Plan - Prasanna has to come up with a precise problem statement for his thesis before the end of March 2006
   
 
= Things to do =
 
= Things to do =
Line 12: Line 72:
 
! Description !! Start date !! Deadline !! Time required !! status
 
! Description !! Start date !! Deadline !! Time required !! status
 
|-
 
|-
| Experiments on Empirical Local Search for the homogeneous PTSP|| 02.11.2005 || 16.11.2005 || ~1 week || In progress
+
| Experiments on Empirical Local Search for the homogeneous PTSP|| 02.11.2005 || 16.11.2005 || ~1 week || Done
 
|-
 
|-
 
| Implementation and Experimentation for heterogenous PTSP|| 15.10.2005 || 25.11.2005 || ~1 week || In progress
 
| Implementation and Experimentation for heterogenous PTSP|| 15.10.2005 || 25.11.2005 || ~1 week || In progress

Latest revision as of 15:41, 21 February 2006

Plan (future work)

Hybrid Algorithms for Stochastic COP

Milestone I : Jan-Mar 2006

  • Incremental Local Optimiztion
  • Finalizing experimentations on Local Search (For Journal)
  • Paper submission to ANTS 2006

Milestone II :Mar-May 2006

  • Ideas - Boosted Sampling, Consensus, Expectation & Consencus
  • ACO/F-Race plus Local search experimentations (For Journal)
  • Paper submission to Hybrid Metaheuristics

May - Jun 2006

  • Vacation

Milestone III :July-Oct 2006

  • Transferring ideas from Simulation Literature mainly from Ranking and Selection
  • Priliminary work on hybridising exact and local search techniques

Milestone IV :Nov-Dec 2006

  • Journal paper (EJOR)
  • Paper submission to GECCO 2007

Milestone V :Jan-April 2007

  • Visiting Gutjahr's lab
  • Goal: Hybridisation, both from simulation and exact techniques prespective

Milestone VI :May-June 2007

  • Priliminary results - Conference paper, Advance results - Journal article (Results from the work done at Vienna)

Milestone VII :July-Oct 2007

  • Putting together the results of two hybridisation
  • Min 1 to Max 2 Journal papers

Milestone VIII :Nov 2007 - Mar 2008

  • PhD thesis

Goals

Weekly Meeting

21-02-2006

1) Experimental results - Incremental Local search

Plots should be generated for the following three cases

a) x- axis - iterations, y-axis - distance between two solutions which are randomly generated. b) x- axis -iterations, y-axis distance between the global best solution to the greedy solution (Two graphs, one before the local search and one after the local search) c) x-axis -iterations, y -axis distance between the new global best and the old global best

The structure of the experimental setting and results

a)We will show reinit with time as the stopping criterion is not promising b)We will show reinit with same number of iteration doesn't help

c)Then we will show the experiments on distances d)Additionally, we will show the results when alpha=0


2)Discussion - Structure of the paper ANTS-2006

3)Future Plan - Prasanna has to come up with a precise problem statement for his thesis before the end of March 2006

Things to do

Description Start date Deadline Time required status
Experiments on Empirical Local Search for the homogeneous PTSP 02.11.2005 16.11.2005 ~1 week Done
Implementation and Experimentation for heterogenous PTSP 15.10.2005 25.11.2005 ~1 week In progress
Intergrating Empirical Local Search with ACO/F-Race and Experiments 01.12.2005 01.01.2005 ~1 week To be started after the previous tasks
Experiments 01.12.2005 01.01.2005 ~2 week To be started after the previous tasks

Papers to write

Title Journal/Conference targeted Start date Submission deadline
Empirical Local Search European Journal of Operational Research N.A. N.A.
ACO/F-Race and Empirical Local Search IEEE System, Man and Cybernetics N.A N.A