Hybrid simulation of evolution dynamics in the neural networks basis

Abstract

Conceptual solutions for constructing a hybrid simulation (HS) system for interpreting behavior of the marine dynamic objects (MDO) in an evolving medium are discussed. The analysis is carried out in the functional spaces of behavior and control of the modern catastrophe theory (MCT). The theoretical basis for the implementation of the multifunctional model of the interpretation of HS determines the principle of structural and parametric synthesis of neurodynamic models under conditions of uncertainty based on the hierarchical structure of the software package. The analysis of behavior MDO and the choice of a solution in a neurodynamic environment is carried out on the basis of ensemble forecast and a matrix of strategic decisions. The interpretation of MDO behavior in HS systems on the base neural-fuzzy and multi-agent environment (MMS) is carried out as part of the Big Data processing strategy for large volumes of data un the regime of urgent computing (UC).