On Convergence of Evolutionary Algorithms Powered by Non-random Generators

Abstract

Inherent part of evolutionary algorithms that are based on Darwin theory of evolution and Mendel theory of genetic heritage, are random processes that are used in every evolutionary algorithm like genetic algorithms etc. In this paper we present experiments (based on our previous) of selected evolutionary algorithms and test functions demonstrating impact of non-random generators on performance of the evolutionary algorithms. In our experiments we used differential evolution and SOMA algorithms with functions Griewangk and Rastrigin. We use n periodical deterministic processes (based on deterministic chaos principles) instead of pseudorandom number generators and compare performance of evolutionary algorithms powered by those processes and by pseudorandom number generators. Results presented here has to be understand like numerical demonstration rather than mathematical proofs. Our results (reported sooner and here) suggest hypothesis that certain class of deterministic processes can be used instead of random number generators without lowering the performance of evolutionary algorithms.