Algorithms for cluster-identification of emergency situations in the management of technogenic and fire safety processes

The problem of cluster-identification of emergency situations in the management of the processes of ensuring technogenic and fire safety is considered. The assignment of a situation to a class is carried out by comparing it with typical elements of different classes and selecting the nearest one. To do this, a measure of class proximity is introduced, which depends on the form in which the signs of situations are set. In the case where these features are expressed in deterministic quantities, the square of the Euclidean distance between the vectors of feature values is used as a measure of the proximity of situations (the smaller the distance, the closer the situations are). The corresponding definition of the features of a typical situation is the arithmetic mean of the features in the sample representing the class of the situation. When the features are set by probabilistic values, the measure of proximity is the generalized probability of identifying threatening, critical, and catastrophic situations. In the case when the signs of a situation are set on conceptual scales, it is proposed to use the apparatus of semantic networks to solve the problem, and the process of identifying situations is understood as a multi-step process that includes: a) conceptualization of the problem; b) generation of solution options; c) evaluation and ranking of solutions; d) selection of the preferred solution. This understanding of the process of identifying situations most fully reflects the structure of human intellectual activity and allows us to proceed to the formalization of these operations based on the use of the apparatus of semantic networks.

1. Averkin A.N., Batyrshin I.Z. and others. Fuzzy sets in control and artificial intelligence models. Ed. by D.A. Pospelov. M.: Nauka. 1986:312.

2. Arkhipova N.I., Kulba V.V. Management in emergency situations. Moscow: RSUH, 2008:480.

3. Classification and cluster. Ed. by J. Van Raizin. M.: Mir. 1980:390.

4. Mandel I.D. Cluster analysis. M.: Finance and Statistics. 1988:176.

5. Kochedykov S.S., Noev A. N., Orlova D.E. Deterministic methods of diagnosing crisis situations in critical control systems. Bulletin of the Voronezh Institute of the Federal Penitentiary Service of Russia. 2017;1:86-89.

6. Orlova D.E., Padalko A.V. The use of the apparatus of semantic networks for intellectual decision-making support. Bulletin of the Voronezh Institute of High Technologies. 2021;1(36):61-66.

7. Orlova D.E., Fursov I.V. Probabilistic-statistical algorithm for identifying emergency situations. Bulletin of the Voronezh Institute of High Technologies. 2021;2(37):56-64.

8. Khaydukov D.S. Application of cluster analysis in public administration. Philosophy of Mathematics: actual problems. Moscow: MAKS-Press. 2009:287.

9. Eddous M., Stansfield R. Methods of decision-making. Trans. from English. M.: Audit. 1997:590.

10. Gabrichidze T.G., Fomin P.M., Yannikov I.M. Creation of a territorial complex multi-stage system for monitoring and forecasting emergency situations of a natural, technogenic and biologosocial nature on the territory of the Samara region. Bulletin of the Samara State Aerospace University. Academician S.P. Korolev. 2008;2(15):272-280.