Absolute and delay-dependent stability of equations with a distributed delay (Q423853)

The paper is devoted to the study of the following equation with a distributed delay \[ x'(t) = r(t)\left( \int_{h(t)}^t f(x(s)) d_s R(t,s) - x(t)\right), \;t \geq 0, \tag{1} \] where \(\int_{h(t)}^t d_s R(t,s) = 1,\) the functions \(h\) and \(r\) are Lebesgue measurable and \(h(t) \leq t\), \(\lim_{t \to \infty} h(t) = \infty\), \(r(t) \geq 0\), \(\int_{0}^{\infty} r(s)\,ds = \infty.\) The continuous function \(f\) satisfies \(f(x) > x\) for \(0 < x < K\) and \(0 < f(x) < x\) for \(x > K,\) \(K = \text{const} > 0\). Equation (1) has a unique positive equilibrium \(x(t) \equiv K\). The authors prove that, if all positive solutions of the difference equation \(x_{n+1} = f(x_n)\) tend to \(K\), then any positive solution of (1) converges to \(K\) (independently of the value of the delay). Also, delay-dependent conditions for the global attractivity of \(K\) are obtained.