Averaging principle for functional-differential equations with impulses at variable times via Kurzweil equations. (Q475151): Difference between revisions

The authors consider impulsive functional-differential equations of the form \[ \begin{aligned} y'(t)&=f(y_t,{t/\varepsilon }),\quad \text{ for } t\neq \tau _k(y(t)),\;k\in \mathbb N,\\ \Delta y(t)&=\varepsilon I_k(y(t)),\quad \text{ for } t=\tau _k(y(t)),\;k\in \mathbb N,\tag{1}\end{aligned} \] where \(y\) and \(f\) take values in \(\mathbb R^n\), the functions \(\tau _k:\mathbb R^n\to \mathbb R\) correspond to hypersurfaces in \(\mathbb R^n\), and the impulses described by the operators \(I_k\: \mathbb R^n\to \mathbb R^n\) take place whenever the solution \(y\) intersects one of the hypersurfaces. The main result is an averaging theorem for (1), which claims that the solutions can be approximated by solutions of a certain autonomous functional-differential equation whose right-hand side is obtained by averaging the original right-hand side. The basic idea of the proof is to convert (1) to an equivalent abstract generalized ordinary differential equation. Reviewer's remark: Although the authors promise to consider equations of the form (1) with state-dependent impulses, it seems that their method is applicable only to equations with preassigned moments of impulse action of the form \[ \begin{aligned} y'(t)&=f(y_t,{t/\varepsilon }),\quad \text{ for }t\neq t_k^i,\;k\in \mathbb N,\;i\in \{1,\ldots ,m(\tau _k)\},\\ \Delta y(t)&=\varepsilon I_k(y(t_k)),\quad \hbox { for }t=t_k^i,\;k\in \mathbb N,\;i\in \{1,\ldots ,m(\tau _k)\}.\tag{2}\end{aligned} \] The reason is that in (1), the moments \(t_k^i\) at which the solution \(y\) intersects the hypersurface given by \(\tau _k\) depend on the choice of \(y\). However, \(t_k^i\) appear in the definition of the function \(h\) on page 1301, while the existing theory of generalized differential equations requires that \(h\) does not depend on \(y\). In other words, the proofs are correct only for (2), where the moments \(t_k^i\) are the same for all solutions; this is the same error as in the earlier papers [\textit{S. M. Afonso et al.}, Math. Nachr. 285, No. 5--6, 545--561 (2012; Zbl 1252.34076)] and [\textit{S. M. Afonso et al.}, Bull. Sci. Math. 137, No. 2, 189--214 (2013; Zbl 1273.34085)]. Note that an averaging theorem for functional-differential equations without impulses was obtained in the paper [the authors, J. Math. Anal. Appl. 382, No. 1, 77--85 (2011; Zbl 1226.34075)].