Inner-outer factorization of nonlinear operators (Q1187777): Difference between revisions

This article deals with the stable discrete nonlinear system \[ x_{n+1}=F(x_ n,u_ n),\quad y_ n=G(x_ n,u_ n), \] where \(x_ n\) are successive states of the system from the state space \(X=\mathbb{R}^ d\), \(u_ n\) and \(y_ n\) are successive inputs and outputs of the system from the input and output spaces \(U=Y=\mathbb{R}^ n\), \(F\) and \(G\) are given smooth functions. Under some natural assumptions, the authors describe the input-output mapping \(M\) and study the possibility of its representation as a convolution \(M=\theta\circ Q\) where \(\theta\) is ``energy conserving'' and \(Q\) is a stable with stable inverse; such representations are an analogue for the inner-outer-factorization of matrix-valued analytic functions on the unit disc. Some explanatory examples are presented as well.