Genetic Algorithm: from Hegemony to Chaos

Description

Genetic algorithms are known to be convergent algorithms, with final population tending to become homogeneous. In this paper we show that it is possible to exhibit complex dynamics for a genetic algorithm by slightly modifying the canonical algorithm. Indeed, adding a meta level in the interpretation of the individuals, associated with a coupling between indivuduals gives rise to periodic or chaotic behavior.

Abstract

International audience

Additional details