by Wojciech Jaśkowski, Bartosz Wieloch, Krzysztof Krawiec
Abstract:
We introduce fitnessless coevolution (FC), a novel method of comparative one-population coevolution. FC plays games between individuals to settle tournaments in the selection phase and skips the typical phase of evaluation. The selection operator applies a single-elimination tournament to a randomly drawn group of individuals, and the winner of the final round becomes the result of selection. Therefore, FC does not involve explicit fitness measure. We prove that, under a condition of transitivity of the payoff matrix, the dynamics of FC is identical to that of the traditional evolutionary algorithm. The experimental results, obtained on a diversified group of problems, demonstrate that FC is able to produce solutions that are equally good or better than solutions obtained using fitness-based one-population coevolution with different selection methods.
Reference:
Fitnessless Coevolution (Wojciech Jaśkowski, Bartosz Wieloch, Krzysztof Krawiec), In GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation (Maarten Keijzer, ed.), Association for Computing Machinery, 2008.
Bibtex Entry:
@InProceedings{Jaskowski2008fitnessless, Title = {Fitnessless Coevolution}, Author = {Wojciech Jaśkowski and Bartosz Wieloch and Krzysztof Krawiec}, Booktitle = {GECCO '08: Proceedings of the 10th annual conference on Genetic and evolutionary computation}, Year = {2008}, Address = {Atlanta, GA, USA}, Editor = {Maarten Keijzer}, Month = {jul}, Organization = {Association for Computing Machinery}, Pages = {355--362}, Publisher = {Association for Computing Machinery}, Abstract = {We introduce fitnessless coevolution (FC), a novel method of comparative one-population coevolution. FC plays games between individuals to settle tournaments in the selection phase and skips the typical phase of evaluation. The selection operator applies a single-elimination tournament to a randomly drawn group of individuals, and the winner of the final round becomes the result of selection. Therefore, FC does not involve explicit fitness measure. We prove that, under a condition of transitivity of the payoff matrix, the dynamics of FC is identical to that of the traditional evolutionary algorithm. The experimental results, obtained on a diversified group of problems, demonstrate that FC is able to produce solutions that are equally good or better than solutions obtained using fitness-based one-population coevolution with different selection methods.}, Keywords = {games, one-population coevolution, selection methods, tic tac toe, nim, rosenbrock, rastrigin, coevolutionary algorithm, genetic programming, coevolutionary dynamics}, Url = {http://www.cs.put.poznan.pl/wjaskowski/pub/papers/jaskowski08fitnessless.pdf} }