Influence of Experiment Design in GPA Investigating with Respect to PRNGs

This paper analyses the influence of experiment parameters onto the reliability of experiments with genetic programming algorithms. The paper is focused on the required number of experiments and especially on the influence of parallel execution which affect not only the order of thread execution but...

Full description

Saved in:
Bibliographic Details
Published in:Mendel (Brno (Czech Republic)) Vol. 24; no. 2
Main Author: Brandejsky, Tomas
Format: Journal Article
Language:English
Published: Brno University of Technology 21.12.2018
Subjects:
Efficiency
Experiment Repeatability
Experiment Results Reliability
Genetic Programming Algorithm
Number of Experiments
Pseudo Random Number Generator
ISSN:1803-3814, 2571-3701
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper analyses the influence of experiment parameters onto the reliability of experiments with genetic programming algorithms. The paper is focused on the required number of experiments and especially on the influence of parallel execution which affect not only the order of thread execution but also behaviors of pseudo random number generators, which frequently do not respect recommendation of C++11 standard and are not implemented as thread safe. The observations and the effect of the suggested improvements are demonstrated on results of 720,000 experiments.
ISSN:1803-3814
2571-3701
DOI:10.13164/mendel.2018.2.069