Voskoboynikova G.M.   Khairetdinov M.S.  

Posteriori Algorithms in the Inverse Problems of Geophysics

Reporter: Voskoboynikova G.M.

At present, the monitoring, prediction, and prevention of natural and technogenic disasters are among priority
problems. Many of them are associated with geophysical monitoring of natural and technogenic events and the
preceding geodynamic processes. Such events are earthquakes, volcano eruptions, lunar and solar tides, landslides,
falls of celestial bodies, quarry explosions causing technogenic earthquakes, etc. Monitoring has several successive
stages, including recording of responses to events and measurement of their major parameters, such as travel times of seismic waves or initial waveforms. At the final stage, inverse problems of determining the geographical location and recording time of an event are solved. The problem of determining the geometric parameters of the underground
zone of preparation for catastrophic events is even more complicated. A popular method for solving the inverse
problems is the least-squares method based on simple calculations. At the same time, this method is sensitive to
crude measurement errors (large deviations) in the initial data, which indicates its limited character [1]. Therefore,
it is important to increase the accuracy of estimating the wave parameters in noise. In this paper, a new approach
is proposed. In comparison to the known methods of statistical data processing, it provides higher accuracy in
the measurement of wave arrival times and simultaneous separation of their forms. This approach is based on a
posteriori computational algorithms of discrete optimization. The results of numerical experiments on estimating
the accuracy and noise immunity of the algorithms are presented

To reports list