On 2005-10-14 at 16:30:00 (Brussels Time) |
Abstract
The concept of "intrinsic emergence" offers a clever way to reduce the search space cardinality and then improves the convergence to the solution. So, a second search process has to be engaged in the space of the observables and two Simple Genetic Algorithms are intertwined to solve the whole problem : one in the original space and one in the space of observables of the original one. After an intuitive application to a cellular automata, we have extended the algorithm to all optimisation problems which can be represented by bit string chromosome: observers are represented by groups of given loci where the genes take the same allele. To test its efficiency, the algorithm is applied on hard problem for genetic algorithms (GAs): hierarchichal problems and in particular Royal Road functions and Hierarchical-if-and-only-if (HIFF) function. The results are compared to those obtained with other algorithms.
Keywords
Intrinsic Emergence, Genetic Algorithm, Royal Road functions, Hierarchical problems, metaheuristics
References
-
Hugues Bersini. (2004)
Whatever emerges should be intrinsically useful.
In
Artificial Life 9. The MIT Press. pp. 226--231.
-
Christophe Philemotte and Hugues Bersini. (2005)
CoEvolution of Effective Observers and Observed Multi-Agents System.
In
ECAL 2005 - LNAI 3630. Springer Verlag. pp. 785--749.
In press.
See ftp://iridia.ulb.ac.be/pub/cphilemo/preprint-ecal2005.pdf
-
Christophe Philemotte and Hugues Bersini. (2005)
Intrinsic Emergence boosts Adaptive Capacity.
In
GECCO'05: Proceedings of the 2005 conference on Genetic and evolutionary computation. ACM Press, New York, NY, USA. pp. 559--660.
See ftp://iridia.ulb.ac.be/pub/cphilemo/p559.pdf