Fast-slow dynamics in logistic models with slowly varying parameters (Q375586)

Consider the initial value problem \[ {dp\over dt}= r(\varepsilon t)p\Biggl(1- {p\over k(\varepsilon t)}\Biggr)- h(\varepsilon t),\;p(0)= p_0,\tag{\(*\)} \] where \(\varepsilon\) is a small positive parameter, that is, the parameters \(r\) and \(k\) change slowly in times. \((*)\) is equivalent to the singularly perturbed initial value problem \[ \varepsilon{dp\over d\tau}= r(\tau) p\Biggl(1-{p\over k(\tau)}\Biggr)- h(\tau),\;p(0)= p_0. \] Using the method of lower and upper solutions, the authors construct an explicit approximate solution \(p_a(t,\varepsilon)\) of \((*)\) on the finite interval \([0,T]\) such that it holds \[ |p(t,\varepsilon)- p_a(t,\varepsilon)|\leq \gamma\varepsilon\quad\text{for }t\in [0,T], \] where \(\gamma\) is some constant not depending on \(\varepsilon\). A lower bound for \(\gamma\) is given.