Chaotic transport in dynamical systems (Q1189430)

`Transport' is considered to be an abstract notion: the change of position of points in phase space. The analysis is based on qualitative geometric theory of dynamic systems, supplemented by numerical computations. It is tacitly assumed that some phase space points (intrinsically of zero-size) are occupied by physical particles (necessarily of finite size), and that the change of position of the former describes realistically the `transport' of the latter under the influence of `forces' described (necessarily approximately) by the equations of motion. The claim of the interdisciplinary usefulness of such an approach is based on a particular generalization of the notion of separatrix, applicable to the study of steady states and transients. Dynamic `chaos' is handled by statistical methods, and is thus reduced to a minor feature. Equations with quasiperiodic forces are replaced by higher-order autonomous equations amenable to a standard treatment, because in abstract theory the number of dimensions is intrinsically immaterial. Concrete information is obtained from ``slices'' (two- dimensional projections). The presentation is that of an intermediate-level graduate textbook. It includes exercises, but omits most of the difficult proofs. A serious error in the original Mel'nikov argument [\textit{V. K. Mel'nikov}, Trans. Mosc. Math. Soc. 12(1963), 1--56 (1965); translation from Tr. Mosk. Mat. O.-va 12, 3--52 (1963; Zbl 0135.31001)], lost in the Russian-English translation, is not mentioned. (The result is necessarily local and not global, as \textit{S. Lattès} has shown, viz. Ann. Mat. (3) 13, 1--69 (1906; JFM 37.0590.02)). The references are rather selective and reflect only the views of one school of thought. ``Slices'' are a notoriously inefficient means of extracting and conveying information (3-dimensional example: Klein's bottle, a closed surface but not a separatrix). The book contains six chapters of increasing `dimensionality'. It is easy to read for mathematically inclined persons, because methodically mathematical results are transposed into simplified models of nominally applied problems. Not a single `inverse' problem is considered: starting from a genuinely applied set of equations, establish and construct a set of relevant solutions.