Methods for Creating Metamodels: State of the Question
DOI:
https://doi.org/10.31649/1997-9266-2020-151-4-74-88Keywords:
response hypersurface, approximation, resource intensity, metamodel, geometric metamodels, stochastic metamodels, heuristic metamodels, neural networks, additive regression, associative machinesAbstract
There has been performed the generalization of materials of modern research in the field of mathematical The classification was carried out on the basis of the method used to create metamodels. The complexity and feasibility of using various techniques in specific cases were evaluated. Particular attention was paid to the construction of metamodels for multidimensional response hypersurfaces complex in topology. The geometric, stochastic, and heuristic classes of used metamodels were critically considered. The concentrated attention was paid to polynomial and spline-metamodels as to representatives of the class of geometric metamodels. A brief description of the main ideas of their construction, the necessary mathematical apparatus of implementation, lists the disadvantages and advantages of correct practical use in numerical experiments. Similarly, stochastic surrogate models, to which it is advisable to attribute regression models based on Gaussian processes or kriging models and models based on radial basis functions, were considered. In addition, a class of heuristic metamodels, which includes models on artificial neural networks, models using the method of group accounting of arguments and support-vector machines, was considered. Regression models based on radial basis neural networks and multilayer perceptrons were analyzed. The results of theoretical studies on surrogate models using multiple neural networks, that is, associative machines, were generalized and systematized. The features of constructing such machines of a static structure with various methods for obtaining collective coordinated composite of solution networks, in particular, with ensemble averaging and boosting, were given. The effectiveness of increasing the accuracy of approximation capabilities of metamodels using hybrid techniques for the simultaneous use of neural network technologies and additive regression, decomposition of the search area, was noted. According to the results of studies, it was found that for response hypersurfaces of complex topology in order to increase the accuracy of approximation, it makes sense to use a hybrid approach, which consists of the simultaneous application of decomposition technologies of the search area and neural networks built on the techniques of associative machines with various methods for obtaining solutions.
References
A. I. J. Forrester, A. Sóbester, and A. J. Keane, Engineering design via surrogate modelling: a practical guide. Chichester: Wiley, 2008.
Е. В. Бурнаев, и П. В. Приходько, «Методология построения суррогатных моделей для аппроксимации пространственно неоднородных функций,» Труды МФТИ, т. 5, № 4, с. 122-132, 2013.
А. О. Глебов, С. В. Карпов, и С. В. Карпушкин, «Методика оптимизации режимных и конструктивных характеристик нагревательной плиты вулканизационного пресса,» Вестник Тамб. гос. Техн, т. 19, № 1, с. 137-151, 2013.
М. А. Чубань, «Аппроксимация поверхности отклика для использования в процессе параметрического синтеза машиностроительных конструкций,» Вестник Нац. техн. ун-та "ХПИ" , сб. науч. тр. Темат. вып.: Транспортное машиностроение. Харьков : НТУ "ХПИ", т. 43, № 1152, с. 161-164, 2015.
Е. В. Бурнаев, П. Ерофеев, А. Зайцев, Д. Кононенко, и Е. Капушев, «Суррогатное моделирование и оптимизация профиля крыла самолета на основе гауссовских процессов.» [Электронный ресурс]. Режим доступа:
http://itas2012.iitp.ru/pdf/1569602325.pdf. Дата обращения: Ноябрь 04, 2015.
M. R. Garifullin, A. V. Barabash, E. A. Naumova, O. V. Zhuvak, T. Jokinen, and M. Heinisuo, “Surrogate modeling for initial rotational stiffness of welded tubular joints,” Magazine of Civil Engineering, no. 3, рр. 53-76, 2016. https://doi.org/10.5862/MCE.63.4.
S. Koziel, D. Echeverrı´a-Ciaurri, and L. Leifsson, “Surrogate-based methods,” in Computational Optimization Methods and Algorithms. Berlin: Springer-Verlag, 2011, pp. 33-59.
С. Г. Радченко, «Анализ методов моделирования сложных систем,» Математичні машини і системи, № 4, с. 123-127, 2015.
J. Friedman, “Multivariate adaptive regression splines (with discussion),” Annals of Statistics, no. 19, pp. 1-141, 1991.
В. Р. Целых, «Многомерные адаптивные регрессионные сплайны,» Машинное обучение и анализ данных, т. 3, № 1, с. 272-278, 2012.
М. Г. Беляев, «Аппроксимация многомерных зависимостей по структурированным выборкам,» Искусственный интеллект и принятие решений, № 3. с. 24-39, 2013.
David J. C. MacKay, Information Theory, Inference and Learning Algorithms. Cambridge: Cambridge University Press. 2003.
S. Bilicz, M. Lambert, S. Gyimothy, and J. Pavo, “Solution of inverse problems in nondestructive testing by a kriging-based surrogate model,” IEEE Transactions on Magnetics, vol. 48, no. 2, 2012. https://doi.org/10.1109/TMAG.2011.2172196.
Е. В. Бурнаев, М. Панов, Д. Кононенко, и И. Коноваленко, «Сравнительный анализ процедур оптимизации на основе гауссовских процессов.» [Электронный ресурс]. Режим доступа: http://itas2012.iitp.ru/pdf/1569602385.pdf. Дата обращения: Ноябрь 04, 2015.
Е. В. Бурнаев, М. Е. Панов, и А. А. Зайцев, «Регрессия на основе нестационарных гауссовских процессов с байесовской регуляризацией,» Информационные процессы. т. 15, № 3, с. 298-313, 2015.
Е. В. Бурнаев, П. Д. Ерофеев, и П. В. Приходько, «Выделение главных направлений в задаче аппроксимации на основе гауссовских процессов,» Труды МФТИ, т. 5, № 3, с. 24-35, 2013.
H. Fang, and M. F. Horstemeyer, “Global response approximation with radial basis functions,” Engineering optimization, vol. 38, no. 4, pp. 407–424. 2006. https://doi.org/10.1080/03052150500422294.
S. De Marchi, and E. Perracchiono, Lectures on Radial Basis Functions. Preprint, 2018.
Саймон Хайкин, Нейронные сети: полный курс. (2-е изд.) Москва, РФ: Издательский дом «Вильямс», 2006.
П. В. Афонин, «Система оптимизации на основе имитационного моделирования, генетического алгоритма и нейросетевых метамоделей,» на Межд. конф. Knowledge-Dialogue-Solutions, Varna, 2007, с. 60-63.
П. В. Афонин, «Оптимизация моделей сложных систем на основе метаэвристических алгоритмов и нейронных сетей,» Инженерный вестник: электронный научно-технический журнал, т. 11, с. 508–516, 2016.
A. Géron, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow. O'Reilly Media, 2019.
N. Duffy, and D. P. Helmbold, “Boosting Methods for Regression,” Machine Learning. vol. 47, pp. 153-200, 2002. https://doi.org/10.1023/A:1013685603443.
В. П. Боровиков, Нейронные сети. STATISTICA Neural Networks: Методология и технологии современного анализа данных. М., РФ: Горячая Линия-Телеком, 2008.
W. Beyer, M. Liebscher, M. Beer, et al. Neural Network Based Response Surface. Methods – a Comparative Study. LS-DYNA Anwenderforum: Ulm. 2006.
С. В. Ковалевский, и В. Б. Гитис, «Аппроксимация функций с помощью каскадных нейроподобных сетей,» Штучний інтелект, № 4. с. 589-593, 2008.
V. Ya. Halchenko, R. V. Trembovetska, V. V. Tychkov, and A. V. Storchak, “Nonlinear surrogate synthesis of the surface circular eddy current probes,” Przegląd elektrotechniczny, vol. 9, pp. 76-82, 2019. https://doi.org/10.15199/48.2019.09.15.
R. V. Trembovetska, V. Ya. Halchenko, and V. V. Tychkov, “Multiparameter hybrid neural network metamodel of eddy current probes with volumetric structure of excitation system,” International Scientific Journal «Mathematical Modeling», vol. 3, no. 4, pp. 113-116, 2019. [Electronic resource]. Available: https://stumejournals.com/journals/mm/2019/4/113.
V. Ya. Halchenko, R. V. Trembovetska, and V. V. Tychkov, “Development of excitation structure RBF-metamodels of moving concentric eddy current probe,” Electrical engineering & electromechanics, no. 2, pp. 28-38, 2019. https://doi.org/10.20998/2074-272X.2019.2.05.
Х. Бринк, Дж. Ричардс, и М. Феверолф, Машинное обучение. Спб., РФ: Питер, 2017.
А. Г. Ивахненко, Индуктивный метод самоорганизации моделей сложных систем. Киев: Наук. Думка. 1982.
A. G. Ivakhnenko, and G. A. Ivakhnenko, “The Review of Problems Solvable by Algorithms of the Group Method of Data Handling (GMDH),” International Journal of Pattern Recognition and Image Analysis: Advanced in Mathematical Theory and Application, vol. 5, no. 4, pp. 527-535, 1995.
GMDH – General description of the GMDH. [Electronic resource]. Available: http://www.gmdh.net/GMDH_abo.htm . 2014.
GMDH – Spectrum of the GMDH algorithms. [Electronic resource]. Available: http://www.gmdh.net/GMDH_alg.htm . 2014.
F. Parrella, “Online support vector regression.” Thesis Inf. Sci, Dept. of Inf. Sci. Univ. of Genoa, Italy, 2007.
В. Я. Гальченко, Р. В. Трембовецька, і В. В. Тичков, «Застосування нейрокомп’ютинга на етапі побудови метамоделей в процесі оптимального сурогатного синтезу антен,» Вісник НТУУ «КПІ», серія "Радіотехніка". Радіоапаратобудування, № 74, с. 60-72, 2018. https://doi.org/10.20535/RADAP.2018.74.60-72.
Р. В. Трембовецька, В. Я. Гальченко, і В. В. Тичков, «Побудова MLP-метамоделі накладного вихрострумового перетворювача для задач сурогатного оптимального синтезу,» Технічні вісті, № 1 (47), № 2 (48), c. 27-31, 2018. [Електронний ресурс]. Режим доступу: https://drive.google.com/file/d/1WlTMRuV9GsWCBvT3X0JiNWbkhHm1K0mi/view.
Downloads
-
PDF (Українська)
Downloads: 777
Published
How to Cite
Issue
Section
License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
