An equivalence theorem concerning population growth in a variable environment (Q1415129)

The authors analyze the long-time behavior of positive solutions to a certain class of first-order ordinary differential equations that includes a generalization of the Weissing-Huisman equation given by \(\frac{dw}{dt}=r(t)\ln ( \frac{H(t)+I(t)}{H(t)+I(t) \text{ e}^{-k(t)w}}) -l(t)w\), where \(r\), \(H\), \(I\), \(k\), \(l\), are positive continuous functions of \(t\geq 0\) which are bounded above and bounded from away from \(0\), and a generalization of the Turner-Bradley-Kirk-Pruitt equation given by \(\frac{dx}{dt}= c(t)x^{1-np}(k^n-x^n)^{\{1+p\}}+ d(t)x-e(t)x^2\), where \(c(t)>0\), \(e(t)\geq 0\), and \(d\) are continuous functions of \(t\geq 0\), \(n\), \(k>0\), \(p\geq 0\), are constants and \(\{.\}\) designates the generalized power function: \(y^{\{m\}}=(\text{sgn }y)\left| y\right| ^m\). These equations represent important models of growth and competition in mathematical biology. The class of first-order equations that is under discussion here is \(\frac{dx}{dt}=xf(t,x)\), \(x(0)>0\), where \(f(t,x)\) is continuous for \(t\geq 0\) and \(x>0\). This equation was introduced by A. N. Kolmogorov in 1936. The existence of positive solutions \(x(t)\) of the problem above is given by Theorem 7.1: suppose further that there are continuous functions \(g\), \(h\), with \(g(t)>0\), such that \(f(t,\xi )>-h(t)\) for all \(t\), and \(\xi \) with \(0<\xi <g(t)\). Then, given a solution \(x(t)\) of the problem, we have \(x(t)>0\) for as long as \(x(t)\) exists to the right of \(0\). Theorem 2.1 is the main result of the paper: assume that \(f(t,x)\) is continuous for \(t\geq 0\), \(x>0\), and that all solutions of the problem above exist in \([0,+\infty )\) and are positive. Suppose further that there exist an interval \(I\subseteq [0,+\infty )\), finite or not, and a continuous nonnegative function \(\lambda =\lambda (t)\) of variable \(t\) such that \(\frac{\partial f}{\partial x}(t,\xi )\leq -\lambda (t)\), \(t\geq 0\), \(\xi \in I\), and \(\int_0^\infty \lambda (s)\,ds=+\infty \). Finally, consider a solution \(y(t)\) of the problem such that \(y(t)\in I\), \(t\geq 0\), and \(\inf y(t)>0\). Then, for any solution \(x(t)\) of the problem with \(x(t)\in I\) for all \(t\geq 0\), we have \(\inf x(t)>0\) and \(\lim_{t\rightarrow +\infty }\frac{x(t)}{y(t)}=1\). The presence/absence of so-called persistent solutions, that is, solutions \(x(t)\) with \(\inf x(t)>0\), which play an essential role in the discussion, is investigated by averaging of \(f(t,0)\): \(\frac 1T\int_t^{t+T}f(\tau ,0)\,d\tau \geq c>0\) or \(\int_0^\infty f(\tau ,0)\,d\tau =-\infty \). For instance, the second condition, completed with \(f(t,x)\leq f(t,0)\) for \(x>0\), yields the nonexistence of such solutions. The paper is beautifully written and the results are accompanied by interesting examples.