scientific article; zbMATH DE number 1784056

From MaRDI portal
Publication:4546273





The authors propose a method of construction of a minimal realization for completely solvable discrete stationary system. Consider the discrete stationary system \[ \begin{cases} x(k+ e_1)= A_ix(k)+ B_i u(k),\quad &1\leq i\leq m,\\ y(k)= Cx(k)+ Du(k),\quad &k\in\mathbb{Z}^m_+,\end{cases}\tag{S} \] where \(x:\mathbb{Z}^m_+\to E\), \(u: \mathbb{Z}^m_+\to U\), \(y:\mathbb{Z}^m_+\to Y\), \(E\), \(U\), \(Y\) are real or complex vector spaces of finite dimension, \(A_i: E\to E\), \(1\leq i\leq m\), \(B_i: U\to E\), \(1\leq i\leq m\), \(C: E\to Y\), \(D: U\to Y\) are linear operators, \((e_1,\dots, e_m)\) is the canonic basis in \(\mathbb{R}^m\) and \(k= (k_1,\dots, k_m)= \sum^m_{i=1} k_i e_i\in \mathbb{Z}^m_+\). The authors transform the system (S) into a system of minimal dimension such that the vectors \(u\) and \(y\) are the ame for both systems. Denote \((S_1)\), \((S_2)\) the systems of type (S) obtained and \(x^{(1)}_0\), \(x^{(2)}_0\) the corresponding vectors \(x(k)\in E\). If for the vectors \(x^{(1)}_0\), \(x^{(2)}_0\) of systems \((S_1)\), \((S_2)\), the vectors \(u(k)\), \(y(k)\) are the same, then the vectors \(x^{(1)}_0\), \(x^{(2)}_0\) are called equivalent. The authors define a matrix \(K\) such that the vectors \(x^{(1)}_0\), \(x^{(2)}_0\) are equivalent if and only if \(Kx^{(1)}_0= Kx^{(2)}_0\). The system (S) is called minimal if it has not different equivalent vectors \(x(k)\in E\). Let \(\dim E=n\), \(\text{rank\,}K= r< n\), \(F\) be a matrix of dimension \(r\times n\) with \(r\) linearly independent lines of the matrix \(K\), \(FR= I_r\) and \((\overline S)\) the system (S) with matrices \(\overline A_i= FA_iR\), \(\overline B_i= FB_i\), \(1\leq i\leq m\), \(\overline C= C\), \(\overline D= D\), system \((\overline S)\) being minimal with the corresponding matrix \(K\) denoted \(\overline K\). The systems \((S_1)\), \((S_2)\) are called equivalent if to every \(x^{(1)}_0\) correspond at least an equivalent \(x^{(2)}_0\) and conversely. Then the vector \(x(k)\in E\) of system (S) denoted \(x_0\) is equivalent to the vector \(x(k)\in E\) of system \((\overline S)\) is given by \(z_0= Fx_0\) and the vector \(x(k)\in E\) of system \((\overline S)\) denoted \(z_0\) is equivalent to the vector \(x(k)\in E\) of system (S) given by \(x_0= Rz_0\). The system \((S_2)\) is called the minimal realization of system \((S_1)\) if \((S_2)\) is minimal and equivalent to \((S_1)\). The authors prove that the system \((\overline S)\) is a minimal realization of the system (S).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4546273)