CLASSIFICATION OF MACHINE LEARNING ALGORITHMS IN THE FIELD OF CONDUCTING EXERCISES OF FIRE PROTECTION UNITS

Shkitronov Mikhail Evgenievich

The topic relevance is due to the growing need to improve the effectiveness of training of fire protection units by optimizing the planning and conduct of exercises. The research goal is to develop a classification of machine learning algorithms applicable to solving problems that arise in the process of conducting exercises of fire protection units. To achieve this goal, the following tasks were solved: an analysis of the regulatory framework and methodological recommendations for the organization of fire protection exercises was carried out; the key factors of the effectiveness of the exercises were identified. The existing machine learning algorithms are considered in the context of their applicability to solving problems related to planning, modeling and analyzing the results of exercises. The following research methods are used: analysis of scientific literature on the organization of exercises and the use of machine learning, systematization and classification of the data obtained, expert assessment of the applicability of algorithms. Based on the study, a classification of machine learning algorithms is proposed, which makes it possible to optimize the process of choosing the most suitable algorithm for solving a specific task within the framework of fire protection unit exercises. We also revealed that reinforcement learning algorithms have the greatest potential for automating the process of developing exercise scenarios and adapting tactical decisions in real time.

Keywords:machine learning, artificial intelligence algorithms, fire protection, fire extinguishing, innovative technologies