Almost sure exponential stabilization of linear stochastic systems driven by a general $α$-stable process via discrete-time state feedback

Abstract

Given an unstable linear scalar differential equation $dx(t)=θx(t)dt$, $θ>0$, we show that it can be almost surely exponentially stabilized by a discrete-time feedback control driven by a general $α$-stable process. More precisely, consider the stochastically controlled system $dx(t)=θx(t)dt+bx(\lfloor t/τ\rfloorτ)dZ(t)$, where $Z(t)$ is a general $α$-stable process with $α\in(1,2)$, and $\lfloor t/τ\rfloor$ denotes the integer part of $t/τ$. Unlike the extensively studied driven Brownian-driven case, the $α$-stable process has no finite second moment, so high-order moment methods are no longer applicable. By using fractional moments, characteristic functions, and moment inequalities for $α$-stable stochastic integrals, we prove that the stochastically controlled system is almost surely exponentially stable provided that the inequality $|bσ|^α\left(1-β\operatorname{sgn}(b)\tan^2\frac{πα}{2}\right)>(θ+bμ) \frac{-1}{Γ(α)\cos\frac{απ}{2}}$ holds, and that $τ>0$ is sufficiently small.