oa Neural network modeling in the problem of localizing the sources of seismic events in the territory of Ukraine in order to assess the seismic risk zones of industrialized territories.

In order to identify areas of seismic risk in industrial regions of Ukraine, the authors used the possibilities of neural network modeling in the problem of localizing earthquake sources, registered, according to monitoring data 2007–2020, by a network of seismic stations IGF NASU Odessa, Skvira, Poltava, Nikolaev. Local hodographs of P-, S - earthquake waves of the Ukrainian shield and the Dnieper-Donetsk depression in the range of magnitudes 2.7–4.8 were constructed. Modeling localization problems in the operational mode allows to construct with sufficient accuracy the sources of seismic events in the territory of Ukraine, which is confirmed by the examination of the results obtained by global travel time curves D-B. Examples of localization of earthquakes of 2011, 2013 with magnitudes 3.9 and 4.6 in the Krivoy Rog basin area provide additional opportunities for analyzing the structural features of the lithosphere, and in the future – real-time evaluation of the characteristics of the seismic process in the task of its prevention.