Home 
 1. CEC'2002 FDEC Workshop 
 Session 1: Presentations 
 Session 2: Discussions  
 Research Groups Posters 
 2. CEC'2003 FDEC Workshop 
 Session 1: Presentations 
 Session 1: Identified Themes 
 Session 2: Discussions 
 3. CEC'2004 FDEC Workshop 
 Discussions of Themes 
 4. CEC'2005 FDEC Workshop 
 Presentation 
 5. WCCI 2006 ECTC Panel Session 
 White Paper (rev2.5) 
 Presentations 
 Discussions 
 6. CEC'2007 FDEC Workshop 
Session 1: Presentations
Workshop on Future Directions in EC
CEC 2003
10 December 2003, 10 am – 12 noon
Ian: 
Hybridization of EC with others, with hill climbing (at the end of the evolution).
More should be done (not much has been done since the last meeting).
Not enough in making use of the “reproduction” results, e.g., Reproduction of results are difficult from existing publications.
Carlos:
Industries look. Difficulties in Mexico to convince the industries.
Int. Collaborations: Funding is diff. from US, but there are some from south ame., Europe.
EC framework: teaching
MOEA should be in the EC framework. More therectical work, foundation and mathematics, liaise with other OR researchers, more statistics in MOEA. Standard benchmark, public domain software for reproduction
Kalyan:
Important points to ponder in EC research:
  • Know what are EC up against
  • Know their terminologies
  • Know limitations to existing methods (w.r.t. current computing scenario)
  • Know when to use EC
  • Know how to best use EC with existing methods
  • Requires fraternizing with classical groups, not competing with them
Convincing applications:
  • More real world applications
  • Large-scale problem (scalability studies)
  • Problems difficult for existing methods (constraints, discrete restrictions, etc.)
  • Global optimization
  • Other aids: Understanding the problem better
Some potential topics:
  • Given a class of problems, which algorithm (EC, non-EC, hybrid) is best?
  • Which representation-operator combination makes the problem tractable?
  • Target: Epsilon-optimum, what algorithm is best?
  • Online optimization
  • Robust optimization
  • Dynamic optimization
  • EDO and multi-objective optimization
Gary:
  • Importance of industrial application (fast answers to meaningful problems)
  • Hybridization with existing methods
  • Domain expert + EC expert = solution (“bridge building”)
  • Help others to understand the importance of EC to other communities
Industrial application
  • Problems with industry
  • Current solutions: Is the industry satisfied with the current solutions to the problem? Interest in EC is increasing at least within biology
  • Future solution: is EC an answer?
Hybridization
Increases acceptance of approach
There appears to be an increase in hybridization methods in CEC 2003
Bridge building
  • Bridge building is difficult
  • Bioinformatics is an area brimming with
    • Problems
    • Data
    • Knowledgeable people
CEC03 and future
- Increase application,
- Translation to industrial research
- Key hybridization of methods
- Introduction of new approaches from the EC community to other problems.
Remaining problems
- Lack of statistical analysis/multiple sampling
- Lack of parameter tuning
- Open to new EC approaches... any approach that solves the problem the Best,
  • Building better bridges
  • Looking at the bridge from both sides
  • Interpreting how the EC community envisions bridge building
 
Xin Yao:
  • We must have some key and solid applications to be growth. Need more people from industries. (I have a problem, I cannot solve it, can EC solve it). How to get these problems?:
  • We need to understand why and how it is working. One approach: computational time complexity (scalability). For large problems, there is no answer from the EC communities.
Hussesin:
Briefing on the UNSW EC research group.
Industry applications: Need a prototype
Teaching – problems as students don’t have the proper background. Difficult to convince them the use of heuristics in EC
Questions:
Popularities are increased, but could it be a premature convergence?
Gary: Reviewers need to contribute in terms of quality of the papers.
Top