No label defined (Q4490848)

The author's restriction to discrete-time models originates from several considerations: many real systems are of discrete-time nature, often models have to be investigated further on a computer and -- the author's main motivation -- model-based computer control has been for long an important topic and challenge in industry and engineering research. The book starts with a discussion of possible approaches to this modelling task. This introduction is at the same time a motivation to the following five main chapters, which have as background two fundamental questions. First, which class of model structures are available to model a specific system in view of a given task, and second, what types of qualitative behaviour can a certain model class exhibit. Basic are ARMAX models (autoregressive moving average models) and NARMAX-models (nonlinear ARMAX models) together with extensions or variations of these models which are necessary to model also such effects as thresholds, special types of nonlinearities or other more special properties of the system to be modelled. The author presents both, an excellent overview on various types of available mathematical models for discrete-time systems and an in depth investigation of a model's properties such as being linear, (positive) homogeneous, additive, bilinear, nonlinear (exponential or BIBO-)stable etc. A careful discussion of relationships between different model classes together with a chapter on the art of model development complete the ideas and results of the first six chapters. The material is presented with the necessary mathematical rigour and so that also engineers and scientists from various areas of application have easy access. Well chosen examples from various applied fields facilitate the understanding of the main concepts and ideas behind them. Hints to the many references contained in the bibliography facilitate more detailed investigations of the reader. Moreover, the author succeeded with his attempt to make each of the eight chapters reasonably self-contained with frequent cross-references to related topics in other chapters. Thus, teachers at universities will find in this volume a basis for interesting courses for advanced undergraduate and/or for graduate students whereas researchers will find stimulating ideas for both, concrete modelling tasks and theoretical investigations of model types and their properties.