Direct Statistical Estimation of GA Landscape Properties

Colin R. Reeves C.Reeves@coventry.ac.uk    School of Mathematical and Information Sciences Coventry University UK

Abstract

A variety of predictive measures have been suggested for assessing how difficult it might be to solve a particular problem instance using a particular algorithm. However, most of these measures have been indirect. For neighbourhood search methods, one direct indicator of problem difficulty is the number of local optima that exist in the problem landscape. In the case of evolutionary algorithms, the concept of a local optimum is not easy to define, but it is known that GA populations, for example, commonly converge to fixed points or ‘attractors’. Whether we speak of local ...

Get Foundations of Genetic Algorithms 2001 (FOGA 6) now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.