Numerical research of the concrete dynamic systems by methods of pattern recognition and statistical modeling
Larisa G. Teklina, Yuri I. Neimark, Irina S. Gel'fer, Igor V. Kotel'nikov
In this work the new approach to numerical research of the concrete multidimensional dynamic systems on the basis of use of the pattern recognition methods and the statistical modeling is submitted. This approach allows to overcome the difficulties connected with the high dimensions of phase space and space of parameters of researched system. The offered approach is based on consideration not ideal but the real and computable mathematical model having the physical and technical importance. The concept of a coarsened computer phase portrait for dynamic system is entered, and the methods of the coarsened phase portrait construction are offered. The potentialities of the new approach are demonstrated on the example of mathematical model for the immune response of a organism to an infection invasion.