On exponential stability of nonlinear time-varying differential equations (Q1301460): Difference between revisions

The main result of the paper proclaims the following sufficient conditions for exponential stability of the zero solution of the system \(\dot x=f(t,x)\). There exist positive numbers \(T\), \(\nu\), \(\lambda_{\min}\leq\lambda_{\max}\), a sequence \(\{t_{k}\}\), \(t_{k}\to+\infty\) as \(k\to+\infty\), \(t_{k}\to-\infty\) as \(k\to-\infty\), \(t_{k+1}-t_{k}\leq T\) for any integer \(k\), and a function \(V(t,x)\) of Lyapunov type such that \(\lambda_{\min}\|x\|^{2}\leq V(t,x)\leq\lambda_{\max}\|x\|^{2}\), and \(V(t_{k+1},x(t_{k+1}))-V(t_{k},x(t_{k})) \leq-\nu\|x(t_{k})\|^{2}\) for every solution \(x(t)\) of the system. It is clear that we need some additional tools to verify the last condition. The authors use the converse Lyapunov theorem to obtain some results on exponential stability of the zero solution of the system in standard form \(\dot x=\varepsilon f(t,x)\), where \(\varepsilon\) is a small nonnegative parameter, when the zero solution of the averaged system \(\dot x =\varepsilon\overline f(x)\), \(\overline f(x)=\lim_{T\to\infty}(1/2T)\int_{-T}^{T} f(t,x) dt\), is exponentially stable. The authors use the same approach to obtain sufficient conditions on uniform asymptotic stability in SIAM J. Control Optimization 37, No. 4, 997-1010 (1999; Zbl 0924.34047).