Multi-leader-follower group consensus of stochastic time-delay multi-agent systems subject to Markov switching topology (Q2688364)

The paper considers structures of ``agents'' with controlled stochastically perturbed delay dynamics as below \[ \dot{x}_i(t)=(Ax_i(t) + A_\tau x_i(t-\tau) + Bu_i(t))dt + Cx_i(t)d\omega(t),\ i\in N_1\tag{1} \] with \(\omega(t)\) -- a standard Brownian motion. These ``agents'' are called ``followers'' of other ``agents'' called ``leaders'', described by \[ \dot{x}_i(t) = Ax_i(t) + A_\tau x_i(t-\tau) + Bf_i(x_i(t),t) + Cx_i(t)\eta(t),\ i\in N_2\tag{2} \] with \(|f_i(x_i,t)|_\infty\leq\varepsilon\) -- a pre-designed control input. Connections of the equations of the ``agents'' are defined as certain kinds of graphs. It is proposed the following distributed control structure for the ``followers'' \[ u_i(t) = \alpha K_{s_i}\xi_i(t) + \beta \operatorname{sgn}(K_{s_i}\xi_i(t)),\ i\in N_1\tag{3} \] where \[ \xi_i(t) = \sum_{j\in N_1\cup N_2}a_{ij}^{s_i}(x_i(t)-x_j(t)) \] and it is proven that if the closed loop system (1)--(3) is stable in the mean square, then the states of the followers converge to a convex hull generated by the leaders, in the mean square.