Stochastic sensitivity analysis: theory and numerical algorithms

International Conference "Structural and Phase Transformations in Materials: Theory, Computer Modelling and Experiment"

材料科学;计算机科学

Bashkirtseva, I.^1

Ural Federal University, Lenina 51, Ekaterinburg

620000, Russia^1

关键词: Confidence domain;    Constructive methods;    Numerical algorithms;    Probabilistic distribution;    Quasi potentials;    Random disturbances;    Stochastic sensitivity analysis;    Stochastic sensitivity functions;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/192/1/012024/pdf
DOI  :  10.1088/1757-899X/192/1/012024

来源: IOP

【 摘 要 】

A problem of the analysis of nonlinear dynamic systems forced by the random disturbances is considered. For the description of probabilistic distributions around the deterministic attractors, a theory of the quasipotential and its approximations based on the stochastic sensitivity functions (SSF) technique is suggested. Constructive methods for the computation of SSF for the equilibria and limit cycles are presented and discussed. We show how SSF can be used for the visual description of probabilistic distributions in the form of confidence domains.