DocumentCode :
296196
Title :
A non-linearity measure of a problem´s crossover suitability
Author :
Mason, Andrew
Volume :
1
fYear :
1995
fDate :
Nov. 29 1995-Dec. 1 1995
Firstpage :
68
Abstract :
For any problem with an unknown fitness function, justification of an evolutionary algorithm as a research method necessarily relies upon conjectures about that fitness function. This paper formulates apparent crossover partition coefficients (a generalisation of Walsh transforms) and uses these to develop a new model of crossover non-linearity ratios. Experimental runs demonstrate that this theory can offer insights into the apparent tractability of problems under crossover
Keywords :
Degradation; Encoding; Genetic algorithms; Search methods; Trademarks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Evolutionary Computation, 1995., IEEE International Conference on
Conference_Location :
Perth, WA, Australia
Print_ISBN :
0-7803-2759-4
Type :
conf
DOI :
10.1109/ICEC.1995.489121
Filename :
489121
Link To Document :
بازگشت