IEEE CIS Task Force on Multi-Objective Evolutionary
Algorithms
Active Members of the Task Force Group (to be completed)
- Sanaz Mostaghim
(chair)
- Marde Helbig (vice chair)
- Rui Wang (vice
chair)
- Gideon Avigad
- Xinye Cai
- Wei-neng Cheng
- Carlos Coello Coello
- Kalyanmoy Deb
- Juan J. Durillo
- Maoguo Gong
- Hisao Ishibuchi
- Yaochu Jin
- Li Ke
- Hongtao Lei
- Hui Li
- Hong Li
- Jian-Ping Li
- Hailin Liu
- Yajie Liu
- Kaisa Miettinen
- Massimiliano Vasile
- Yong Wang
- Yuping Wang
- Jian Xiong
- Qingfu Zhang
- Jun Zhang
- Tao Zhang
- Aimin Zhou
Why a Tasking Force Group?
The main aim of the task force groups of the IEEE CIS Technical Committee on
Evolutionary Computation is to promote the research in evolutionary
multi-objective Optimization (EMO). EMO is already an established field and
there are many conferences, workshops and special sessions addressing
algorithmic parts of EMO. Our major goals are to focus more on the new aspects
such as theory of EMO, many-objective Optimization, dynamic EMO, Robustness in
EMO, combinations of EMO and MCDM approaches and last but not least
applications.
You are more than welcome to contribute!
What is Multi-Objective Optimization?
Multi-objective optimization refers to the solution of problems with two or
more objectives to be satisfied simultaneously. Normally, such objectives are
in conflict with each other and are expressed in different units. Because of
their nature, multi-objective optimization problems normally have not one but a
set of solutions, which are called Pareto optimal solutions. When such
solutions are plotted in objective function space, the graph produced is called
the Pareto front of the problem.
Why Evolutionary Multi-Objective Algorithms?
Despite the existence of numerous mathematical programming techniques for
multi-objective optimization, evolutionary algorithms are particularly suitable
for these problems because of several reasons:
- Evolutionary
algorithms are less susceptible to the shape or continuity of the Pareto
front, whereas many mathematical programming techniques rely on some a
priori knowledge about such shape.
- Evolutionary
algorithms are population-based. Thus, it is expected that they can
produce several elements of the Pareto optimal set within a single
execution. In contrast, mathematical programming techniques normally
produce a single solution per run.
- Evolutionary
algorithms start with a set of random solutions, whereas mathematical
programming techniques normally require a starting point and the result
that they produce tends to rely on such point.
Where can I start to learn about Multi-Objective
Evolutionary Algorithms?
There are several tutorials available in electronic format. The following is
a representative list:
- The tutorial on
Evolutionary Multiobjective Optimization at CEC
2013, by Kalyanmoy Deb (tutorial slides)
- The tutorial on
Evolutionary Multi-objective Optimization in GECCO 2013, by Dimo Brockhoff (tutorial
slides)
- A Tutorial on
Evolutionary Multi-objective Optimization (2003), by Eckart
Zitzler (tutorial
slides).
Note that there is also a paper version
of this tutorial. The full reference is the following:
Eckart Zitzler, Marco Laumanns and
Stefan Bleuler. A Tutorial on
Evolutionary Multiobjective Optimization, in
Xavier Gandibleux, Marc Sevaux,
Kenneth Sörensen and Vincent T'kindt
(editors), Metaheuristics for Multiobjective
Optimisation, pp. 3--37, Springer. Lecture Notes in
Economics and Mathematical Systems Vol. 535, Berlin, 2004.
- A Short Tutorial
on Evolutionary Multiobjective Optimization
(2001), by Carlos A. Coello Coello
(tutorial
slides).
Note that there is also a paper version
of this tutorial. The full reference is the following:
Carlos A. Coello Coello. A Short Tutorial on Evolutionary Multiobjective
Optimization. In Eckart Zitzler, Kalyanmoy Deb, Lothar Thiele, Carlos A. Coello Coello, and David Corne, editors,
First International Conference on Evolutionary Multi-Criterion Optimization,
pages 21-40. Springer-Verlag. Lecture Notes in
Computer Science No. 1993, 2001.
Additional Resources
If you want to get more in-depth knowledge about multi-objective
evolutionary algorithms, there are two monographs devoted to this topic
currently available:
Books:
- Branke, J., Deb, K., Miettinen, K., Slowinski, R.
(eds.), Multiobjective Optimization: Interactive
and Evolutionary Approaches, Springer-Verlag,
Berlin, Heidelberg, 2008
- Carlos A. Coello Coello, David A. Van Veldhuizen and Gary B. Lamont, Evolutionary
Algorithms for Solving Multi-Objective Problems, Kluwer Academic
Publishers, New York, March 2002, ISBN 0-3064-6762-3.
- Kalyanmoy Deb. Multi-Objective
Optimization using Evolutionary Algorithms, John Wiley & Sons, Chichester, UK, 2001, ISBN 0-471-87339-X.
Webpages:
The EMOO repository contains a lot of
valuable resources for those interested in multi-objective evolutionary
algorithms, such as:
- Over 2500
bibliographic references, many of which are electronically available. This
includes about 150 PhD theses, about 650 journal papers and about 1400
conference papers.
- Contact
information of about 65 researchers working in this area.
- Public-domain
versions of several multi-objective evolutionary algorithms, including
MOGA, MOPSO, NSGA, NSGA-II, microGA, among
others.
- Test functions
and sample Pareto fronts for well-known problems.
- http://www.soft-computing.de/ maintained by Yaochu
Jin, and shares some information on EC methods.
- http://www.macs.hw.ac.uk/~ml355/conference.htm
maintained by Michael A. Lones, and provides a
(almost complete) list of EC related conferences.
EMO Platforms and Programs:
- PISA http://www.tik.ee.ethz.ch/pisa/?page=pisa.php
- JMetal
http://jmetal.sourceforge.net/index.html
- Liger
http://codem.group.shef.ac.uk/index.php/research/software
Multi-Objective Optimization and Games
Here is an interesting application of EMO in Games: Multi-Objective Physical Traveling
Salesman Problem
Events
Some important upcoming events related to multi-objective evolutionary
algorithms are the following:
- IEEE SSCI
2015: Our task force organizes one special session about
Evolutionary Multi-Objective Optimization (Organizers: Sanaz Mostaghim and
Marde Helbig)
- IEEE WCCI2016:
The annual IEEE Congress on Evolutionary Computation -> Our task force
organizes one special session about Evolutionary Multi-Objective
Optimization (Organizers: Sanaz Mostaghim and Kalyanmoy
Deb)
Past Events:
For additional information
If you want to know more about the activities of this working group, or you
want to join us, please contact:
Prof. Dr.-Ing. Sanaz Mostaghim
Faculty of Computer Science
University of Magdeburg
Germany
Sanaz.mostaghim@ovgu.de
Dr. Marde Helbig
Senior Researcher, CSIR: Meraka Institute and
Temporary Senior Lecturer, University of Pretoria
South Africa
mgreeff@googlemail.com
Dr. Rui Wang
Lecturer
College of Information Systems and Management
National University of Defense Technology
China
ruiwangnudt@gmail.com