Nonrandom evolutionary algorithms

Abstract

Inherent part of evolutionary algorithms that are based on Darwin theory of evolution and Mendel theory of genetic heritage, are random processes. In our as well as another researcher papers is successfully discussed possibility to replace pseudorandom number generators by deterministic chaos generator, generating chaos, and then by n periodical series based on deterministic chaos generators. Because all those previous experiments were successful or highly comparable with classical evolutionary algorithms powered by classical pseudorandom number generators, we extend our previous study on use classical mathematical periodical functions (very trivial one) and its use instead of pseudorandom number generators. Results presented here are numerical demonstration (partly of well know facts) rather than mathematical proofs. We open question whether evolutionary algorithms really need pseudorandom number generators or whether (and under what conditions) class of deterministic processes can be used without lowering the performance of evolutionary algorithms.