Stabilization of Volterra equations by noise (Q2498194)

The authors consider the scalar nonlinear convolution Volterra equation \[ x'(t) = - a f(x(t)) + \int_0^t \lambda \exp(- \mu (t-s)) g(x(s)) ds\tag{1} \] and its stochastic counterpart in Itô formulation \[ dX(t) = (- a f(X(t)) + \int_0^t \lambda \exp(- \mu (t-s)) g(X(s)) ds)dt + \sigma X(t) dB(t),\tag{2} \] where \(B\) is a standard Brownian motion. They study the stability of zero solutions of these equations, i.e., exponential asymptotic stability of the zero solution of \((1)\) and almost sure exponential asymptotic stability of the zero solution of \((2)\). The authors provide conditions on the parameters for the latter case to hold. They then discuss regions of the parameter space where the solutions of \((1)\) are unstable, but the zero solution of \((2)\) is still almost surely exponentially asymptotically stable for certain ranges of the parameter \(\sigma\), that is they discuss stabilisation by noise. Examples and extension to more general kernels and systems of equations are also presented.