Ordering Genetic Algorithm Genomes With

Reconstructability Analysis

Stephen Shervais(1) and Martin Zwick(2)

(1) College of Business and Public Administration

Eastern Washington University, Cheney, WA 99004  

(2)  Systems Science Ph.D. Program

Portland State University,Portland, OR  97201  

[sshervais@ewu.edu] (1), [zwickm@pdx.edu] (2)  

 

Keywords: Reconstructability analysis, genetic algorithms, transposition, crossover, optimization, OCCAM

 Abstract

The building block hypothesis implies that genetic algorithm effectiveness is influenced by the relative location of epistatic genes on the chromosome.  We find that this influence exists, but depends on the generation in which it is measured.  Early in the search process it may be more effective to have epistatic genes widely separated.  Late in the search process, effectiveness is improved when they are close together.  The early search effect is weak but still statistically significant; the late search effect is much stronger and plainly visible. We demonstrate both effects with a set of simple problems, and show that information-theoretic reconstructability analysis can be used to decide on optimal gene ordering.

Entire Paper

Discrete Multivariate Modeling Page