Effects of stochastic perturbation on the SIS epidemic system (Q504094): Difference between revisions

In this paper, the authors extend the classical, deterministic SIS model to the following version with stochastic perturbations \[ \begin{aligned} dS(t)&=(\Lambda -\mu S(t)-\beta S(t)I(t)+\lambda I(t))dt+\sigma S(t)dB(t),\\ dI(t)&=(-(\mu +\lambda)I(t)+\beta S(t)I(t))dt+\sigma I(t)dB(t), \end{aligned} \] where \(B(t)\) is a Brownian motion defined on a complete probability space and \(\sigma\) is the variance of the number of potentially infectious contacts that a single infected individual makes with another individual in the small time interval \([t,t+dt)\). As it is customary for disease propagation models, conditions for the extinction and the persistence of the disease are considered. It is shown that if \(\mathcal{R}_\sigma =\frac{\beta \Lambda}{\mu (\mu +\lambda +\frac{1}{2}\sigma ^2)}<1\) \(I(t)\) tends to \(0\) exponentially almost surely and \(S(t)\) admits a unique stationary distribution. If \(\mathcal{R}_\sigma>1\) the disease is shown to be persistent and \((S(t),I(t))\) is observed to possess a unique stationary distribution, for which ergodic properties are established. Consequently, \(\mathcal{R}_\sigma\) plays the role of the basic reproduction number \(\mathcal{R}_0\) of the deterministic counterpart. Since \(\mathcal{R}_\sigma<\mathcal{R}_0\), the condition \(\mathcal{R}_0\leq 1\) is now only a sufficient condition for the extinction of the disease in the stochastic model. Also, the introduction of the stochastic noise is seen to modify the dynamics of the deterministic model. Specifically since \(\lim _{\sigma \to \infty}\mathcal{R}_\sigma =0\), very large perturbations cause the extinction of the disease with probability \(1\). Also since \(\lim _{\sigma \to 0}\mathcal{R}_\sigma =\mathcal{R}_0\), sufficiently small perturbations preserve the properties of \(\mathcal{R}_0\) in a stochastic sense.