Exponential stability of linear delay difference equations with continuous time (Q2351167)

The authors deal with the problem of exponential stability of the linear delay difference equation with continuous time \[ x(t)=\sum_{i=1}^m A_i(t)x(t-h_i)+\int_{-h}^0 B(t,s)x(t+s)\,ds, \quad t\geq t_0\geq 0, \tag{1} \] with \(n\times n\) matrix-valued coefficients \(A_j,B\). One of the main results of the paper reads as follows. { Theorem.} Equation (1) is exponentially stable, provided one of the following conditions holds: {\parindent=6mm \begin{itemize}\item[(i)] There exist \(\alpha>0\) and \(p\in {\mathbb R}^n\), \(p\gg 0\), such that \[ \left(\sum_{i=1}^m|A_i(t)|e^{\alpha h_i}+\int_{-h}^0|B(t,s)|e^{-\alpha s}\,ds\right)p-p\ll 0, \quad t\in {\mathbb R}, \] where the relations \(\ll 0,\gg 0\) mean that all entries of a given vector are negative/positive, and the absolute value of a matrix is the matrix consisting of absolute values of entries of the original matrix. \item[(ii)] There exists \(M\in {\mathbb R}^{n\times n}\) with nonnegative entries such that its spectral radius satisfies \(\rho(M)<1\) and \[ \sum_{i=1}^m|A_i(t)|+\int_{-h}^0|B(t,s)|\,ds \leq M, \quad t\in {\mathbb R}. \] Here the inequality between matrices means that this inequality satisfy all entries. \end{itemize}}