The stability of distributed neutral delay differential systems with Markovian switching (Q2393260)

Let \(r(t)\), \(t\geq 0\), be a right continuous Markov chain taking values in a finite state space \(S= \{1,2,\dots, N\}\), which is independent of the Brownian motion \(w(\cdot)\). In this paper, the authors consider \(n\)-dimensional stochastic neutral distributed delay functional differential system with Markovian switching: \[ \begin{multlined} [x(t)+ g(t, x_t)]= \{(A_r+ \Delta A_r)[x(t)+ g(t, x_t)]+ (A_{1r}+\Delta A_{1r})[x(t- \eta)\\ + g(t- \eta, x_{t-\eta})]\}\,dt +\{(C_r+ \Delta C_r)[x(t)+ g(t, x_t)]\\ + (D_r+ \Delta D_r)[x(t-\eta)+ g(t-\eta, x_{t-\eta})]\}\, dw(t),\end{multlined}\tag{1} \] where \(A_r= A(r(t))\), \(A_{1r}= A_1(r(t))\), \(C_r= C(r(t))\), \(D_r= D(r(t))\) are known constant matrices \(n\times n\) and \(\Delta A_r\), \(\Delta A_1\), \(\Delta C_r\), \(\Delta D_r\) represent the uncertain parts of \(A_r\), \(A_{1r}\), \(C_r\), \(D_r\), respectively. The system (1) is said to exponentially stable in mean square, if \[ \limsup_{t\to\infty}\, {1\over t}\log(\operatorname{E}|x(t, \xi)|^2)< 0, \] where \(x(t,\xi)\) is the solution of (1) with initial data \(x_0=\xi\). The authors discuss the assumptions under which system (1) is exponentially stable in mean square. The results are verified by numerical examples.