Dynamical systems, graphs, and algorithms (Q2431729)

This monograph contains a summary of the author's work on constructive methods for the study of discrete dynamical systems. The set-oriented method is based on a covering of the phase space by finitely many cells. From this covering a directed graph is constructed with vertices corresponding to the cells and edges corresponding to transitions from one cell to another in the original dynamical system. The new dynamical system on the directed graph is called the symbolic image. By refining the covering one hopes to approximate the original system sufficiently well to extract from the symbolic image all information about the localization of periodic orbits, invariant sets or attractors. Moreover, quantities which describe the amount of disorder in the original system, like topological entropy, Lyapunov exponents or the Morse spectrum can also be approximately determined. The structure of the book is very clear with 14 chapters devoted to different dynamical objects such as chain recurrent sets, structural stability or invariant manifolds, followed by two examples: the Ikeda mapping and a discrete food-chain model. Each of the chapters begins with a short but concise introduction to the relevant concepts. After the important objects are defined and their properties are described, their relation to the symbolic image is studied and algorithms are presented. Most of these algorithms are rather simple to describe and can therefore be implemented also by non-specialists or students who want to visualize a given dynamical system and get some idea about its dynamical properties. Numerical analysts may object that there are no error estimates and the complexity of the algorithms does increase heavily from one subdivision step to another even for systems of moderate dimension. Also, for many of the objects studied in the monograph there exist alternative approaches for many years. Although some of them are briefly mentioned in the introduction it would also have been nice to compare convergence and complexity of those algorithms to the ones presented in the book. Despite these minor shortcomings, this monograph is certainly a valuable and very readable reference, in particular for the study of low-dimensional concrete systems with complicated dynamics.