The spectral properties of distributions and asymptotic methods in perturbation theory (Q2901884)

The author investigates the problem of justifying asymptotic expansions for differential equations in standard Bogolyubov form \(D_tx=\varepsilon f(t,x;\varepsilon)\), where \(x(t)\) is an unknown function of the real variable \(t\in\mathbb{R}\) with values in a complex Banach space \(\mathcal B\), \(D_t\) denotes differentiation with respect to \(t\), \(\varepsilon\) is a small real parameter, and \(f\) is a bounded map from \(\mathbb{R}\times {\mathcal B}\times \mathbb{R}\) into \({ \mathcal B}\) which is continuous in \((t,x)\). He uses a modification of the classical Krylov-Bogolyubov method, which allows complications in the construction of higher-order approximations which steam from the `small denominators problem' to be avoided and many of the standard constraints on the bahaviour of the function \(f\) to be eliminated. The approach suggested in this paper is based on some results on the Fourier transforms of distributions.