Towards a theory of practice in metaheuristics design: A machine learning perspective

Issue		RAIRO-Theor. Inf. Appl. Volume 40, Number 2, April-June 2006 Alberto Bertoni: Climbing summits


Page(s)		353 - 369
DOI		10.1051/ita:2006009
Published online		20 July 2006

Theoret. Informatics Appl. 40, 353-369 (2006)
DOI: 10.1051/ita:2006009

Towards a theory of practice in metaheuristics design: A machine learning perspective

Mauro Birattari, Mark Zlochin and Marco Dorigo

IRIDIA, Université Libre de Bruxelles, Brussels, Belgium; mbiro@ulb.ac.be; mzlochin@ulb.ac.be; mdorigo@ulb.ac.be

(Published online 20 July 2006)

Abstract
A number of methodological papers published during the last years testify that a need for a thorough revision of the research methodology is felt by the operations research community - see, for example, [Barr et al., J. Heuristics 1 (1995) 9-32; Eiben and Jelasity, Proceedings of the 2002 Congress on Evolutionary Computation (CEC'2002) 582-587; Hooker, J. Heuristics 1 (1995) 33-42; Rardin and Uzsoy, J. Heuristics 7 (2001) 261-304]. In particular, the performance evaluation of nondeterministic methods, including widely studied metaheuristics such as evolutionary computation and ant colony optimization, requires the definition of new experimental protocols. A careful and thorough analysis of the problem of evaluating metaheuristics reveals strong similarities between this problem and the problem of evaluating learning methods in the machine learning field. In this paper, we show that several conceptual tools commonly used in machine learning - such as, for example, the probabilistic notion of class of instances and the separation between the training and the testing datasets - fit naturally in the context of metaheuristics evaluation. Accordingly, we propose and discuss some principles inspired by the experimental practice in machine learning for guiding the performance evaluation of optimization algorithms. Among these principles, a clear separation between the instances that are used for tuning algorithms and those that are used in the actual evaluation is particularly important for a proper assessment.

What is OpenURL?

The OpenURL standard is a protocol for transmission of metadata describing the resource that you wish to access. An OpenURL link contains article metadata and directs it to the OpenURL server of your choice. The OpenURL server can provide access to the resource and also offer complementary services (specific search engine, export of references...). The OpenURL link can be generated by different means.

If your librarian has set up your subscription with an OpenURL resolver, OpenURL links appear automatically on the abstract pages.
You can define your own OpenURL resolver with your EDPS Account. In this case your choice will be given priority over that of your library.
You can use an add-on for your browser (Firefox or I.E.) to display OpenURL links on a page (see http://www.openly.com/openurlref/). You should disable this module if you wish to use the OpenURL server that you or your library have defined.