2025, Vol. 12, No. 1. - go to content...

Permanent address of this page - https://resources.today/en/25ecor125.html

Метаданные этой статьи так же доступны на русском языке

DOI: 10.15862/25ECOR125 (https://doi.org/10.15862/25ECOR125)

Full article in PDF format (file size: 434.9 KB)

For citation:

Bogonosov K.A. AI decision support systems for choosing the architecture of physical experiments. Russian journal of resources, conservation and recycling. 2025; 12(1). Available at: https://resources.today/PDF/25ECOR125.pdf (in Russian). DOI: 10.15862/25ECOR125

AI decision support systems for choosing the architecture of physical experiments

Bogonosov Konstantin Alexandrovich
Moscow State University of Technology and Management
named after K.G. Razumovsky (First Cossack University), Moscow, Russia
E-mail: k.bogonosov@mgutm.ru

Abstract. The study is devoted to a comprehensive analysis of the development and application of artificial intelligence systems in the field of decision support in the design of the architecture of physical experiments. The paper presents a systematization of modern approaches to the integration of machine learning methods, neural networks and expert systems into the planning and optimization of experimental research in various fields of physical sciences. The analysis demonstrates the transformation of traditional experimental paradigms under the influence of artificial intelligence technologies, which is reflected in the transition from intuitive methods of planning experiments to autonomous intelligent systems capable of independently determining optimal research parameters and adapting to the results in real time. Special attention is paid to the analysis of architectural solutions of modern AI decision support systems, including the integration of high-performance computing platforms, robotic complexes for automated experiments and active learning algorithms to optimize research processes. The paper considers in detail the key technological components of such systems: Bayesian optimization algorithms for targeted exploration of the parameter space, methods for sequential planning experiments taking into account uncertainty, approaches to integrating subject knowledge into machine learning models and technologies for autonomous control of experimental equipment. The study reveals the specifics of various types of physical experiments in terms of architecture requirements for supporting AI systems, including materials science research, optimization of energy systems, development of new compounds, and characterization of physico-chemical properties of substances. The results of the analysis highlight the significant potential of AI systems in accelerating scientific discoveries by automating routine experimental planning processes, improving the accuracy of predicting results and optimizing the use of research resources, while maintaining the critical role of expert knowledge in setting research objectives and interpreting the results.

Keywords: artificial intelligence; decision support systems; architecture of physical experiments; machine learning; Bayesian optimization; automated laboratories; experimental planning

Download article in PDF format