@article {10.3844/jcssp.2019.57.66,
article_type = {journal},
title = {Performance Optimization of Physics Simulations Through Genetic Algorithms},
author = {Shadura, Oksana and Carminati, Federico and Petrenko, Anatoliy},
volume = {15},
number = {1},
year = {2019},
month = {Jan},
pages = {57-66},
doi = {10.3844/jcssp.2019.57.66},
url = {https://thescipub.com/abstract/jcssp.2019.57.66},
abstract = {The GeantV R&D approach is revisiting the standard particle transport simulation approach to be able to benefit from “Single Instruction, Multiple Data” (SIMD) computational architectures or extremely parallel systems like coprocessors and GPUs. The goal of this work is to develop a mechanism for optimizing the programs used for High-Energy Physics (HEP) particle transport simulations using a “black-box” optimization approach. Taking in account that genetic algorithms are among the most widely used “black-box” optimization methods, we analyzed a simplified model that allows precise mathematical definition and description of the genetic algorithm. The work done in this article is focused on the studies of evolutionary algorithms and particularly on stochastic optimization algorithms and unsupervised machine learning methods for the optimization of the parameters of the GeantV applications.},
journal = {Journal of Computer Science},
publisher = {Science Publications}
}