The stability of some class of Itô functional differential equations in the critical case of one zero root (Q1422673): Difference between revisions

The authors consider a system of stochastic differential equations \[ dx=\left(Ax +\sum_{i=1}^kb_if_i\right)dt +\sum_{i=1}^kb_ix^*R_idw,\qquad d\eta= \sum_{i=1}^k\beta_if_idt +\sum_{i=1}^k\beta_ix^*R_idw, \] where \(A\), \(R_i\) are matrices, \(b_i\) are vectors, \(\beta_i\) are scalars (all of them non-random), \(w\) is a Wiener process, and \(f_i(t)\) are real functions depending on the paths of \(\sigma_i=c_i^*x-\rho_i\eta\) in a non-anticipating way; \(c_i\), \(\rho_i\) are non-random vectors and scalars, respectively. Moreover, there is a partition of the time axis \(0=t_{i0} <t_{i1}<t_{i2}<\dots\) also depending on \(\sigma_i\) in a non-anticipative way, such that \(f_i\) is continuous and preserves its sign on each interval \((t_{in},t_{i,n+1})\). The authors claim that these equations describe dynamics of an asynchronous pulse system with \(k\) pulse elements. Under rather lengthy and involved assumptions they prove that the system is stable, i.e., \(\lim_{t\to\infty}E| x(t)| =0\), for any initial conditions \(x(0)\), \(\eta(0)\).