Open AccessMETHOD OF PROCESSING RESULTS OF FACTORY EXPERIMENT
Author(s) -
V. M. Moisyshyn,
М. V. Lyskanych,
Р. А. Жовнірук,
Vasyl Vekeryk,
Y. L. Gavryliv
Publication year - 2021
Publication title -
prikarpatsʹkij vìsnik ntš. čislo
Language(s) - English
Resource type - Journals
ISSN - 2304-7399
DOI - 10.31471/2304-7399-2020-1(59)-44-65
Subject(s) - fractional factorial design , factorial experiment , factorial , design of experiments , product (mathematics) , factory (object oriented programming) , empirical research , mathematics , computer science , main effect , statistics , mathematical analysis , geometry , programming language
Two plans were used: classical and factorial (rational) during experimental research. Classical plan is used when the number of external independent factors is not more than two. Increasing number of external factors leads to using of rational experiment planning which allows to significantly reduce the number of combinations of these factors. For example, with four external factors that change at four levels according to the classical plan number of combinations 44=256. Rational planning of the experiment reduces number of these combinations to 16. To process results of the factorial experiment using two methods proposed by M. M. Protodiakonov and R. S. Yaremiychuk. The article presents a step-by-step method of processing the results of a factorial experiment with an example of simultaneous use of these methods. Comparing the results of processing, the authors believe that the best results are obtained using the Protodiakonov’s method. The authors tested two types of multifactorial empirical model of multiple correlation, one of which is given by the product of partial empirical dependences, the second – by the sum. Using these models there were found that the results of factor experiment processing should be represented by a multifactor empirical model, which is presented by the product of partial empirical dependences.