XPost: comp.ai.genetic, comp.ai.neural-nets, comp.ai.alife
From: a.roli@unich.it
*********************************************************************
CALL FOR PAPERS
Journal of Mathematical Modelling and Algorithms
Special Issue on Hybrid Metaheuristics
EDITORS OF THE SPECIAL ISSUE
Christian Blum, Universitat Politecnica de Catalunya, Barcelona, Spain
Andrea Roli, Universita' "G.D'Annunzio", Pescara, Italy
Michael Sampels, Universite' Libre de Bruxelles, Belgium
INTRODUCTION
Metaheuristics, such as simulated annealing, genetic and evolutionary
algorithms, tabu search, ant colony optimization, scatter search and
iterated local search, have received considerable interest in the
fields of applied artificial intelligence and combinatorial
optimization.
Plenty of hard problems in a huge variety of areas, including
bioinformatics, logistics, engineering, business, etc., have been
tackled successfully with metaheuristic approaches. For many problems
the resulting algorithms are considered to be the state-of-the-art
methods.
For many years the main focus of research was on the application of
single metaheuristics to given problems. In recent years, it has
become evident that the concentration on a sole metaheuristic is
rather restrictive. A skilled combination of concepts of different
metaheuristics, a so called hybrid metaheuristic, can provide a more
efficient behavior and a higher flexibility when dealing with
real-world and large-scale problems.
A quite new field of research is also the hybridization of
metaheuristics with other techniques. Recently, it was observed that
the incorporation of more classical artificial intelligence and
operations research techniques in metaheuristics can be very
beneficial. An example is the use of constraint programming
techniques, such as domain filtering and variable fixing, in order to
shrink the search space that has to be searched by a
metaheuristic. Another example is the incorporation of tree search
procedures in metaheuristics.
The design and implementation of hybrid metaheuristics rises problems
going beyond questions about the design of a single metaheuristic.
Choice and tuning of parameters is for example enlarged by the problem
of how to achieve a proper interaction of different algorithm
components. Interaction can take place at low-level, using functions
from different metaheuristics, but also at high-level, e.g., using a
portfolio of metaheuristics for automated hybridization.
It is implicit with the subject of the special issue that
contributions should address the combination and comparison of
different metaheuristic components and concepts. In contrast to
standard research in metaheuristics, also negative results - e.g., a
component shows poor performance for the majority of test instances -
are of considerable importance in hybridization. Such results have
often been ignored, at least in the publication of results in standard
metaheuristics research.
Further, the above mentioned enlarged selection of parameters will
attract more attention to this part of designing algorithms.
In summary, with this special issue we aim at papers that give good
examples for carefully designed and well-analyzed hybrid
metaheuristics.
The extraction of guidelines for the general design of hybrid
metaheuristics would be desirable.
TOPICS OF INTEREST (NOT LIMITED TO):
* novel combinations of components from different metaheuristics,
* hybridization of metaheuristics and more classical AI and OR
techniques,
* low-level hybridization,
* high-level hybridization, portfolio techniques, expert systems,
* co-operative search,
* taxonomy, terminology, classification of hybrid metaheuristics,
* population-based techniques,
* co-evolution techniques,
* advanced local-search and gradient techniques,
* automated parameter tuning,
* empirical and statistical comparison,
* theoretic aspects of hybridization,
* parallelization,
* software libraries.
IMPORTANT REMARK
Special attention will be given to the validity of results of
experimental work. Decisions in the design of hybrid techniques and
results of the own work in comparison with the prior art are to be
derived using adequate statistical methods.
PAPER SUBMISSION
Researchers are invited to submit papers that make a significant
contribution to the field and that are ideally be between 10 and 20
pages long to hm2004@iridia.ulb.ac.be .
Please use the subject "Special Issue Paper Submission". Every paper
will be reviewed by at least two reviewers. Source code and problem
instances should (if relevant) be made available on the Internet. It
is recommended to use the Kluver Journal Style Files
(http://www.wkap.nl/authors/jrnlstylefiles) when preparing the
submission.
IMPORTANT DATES
Submission of papers: November 30, 2004
Notification of acceptance: March 2005
Camera-ready papers: to be announced
Publication of the Special Issue: Issue 3, 2005
CFP homepage: http://iridia.ulb.ac.be/~hm2004/jmma.html
Journal of Mathematical Modelling and Algorithms homepage:
http://www.kluweronline.com/issn/1570-1166/
[ comp.ai is moderated. To submit, just post and be patient, or if ]
[ that fails mail your article to , and ]
[ ask your news administrator to fix the problems with your system. ]
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
