? J. Andersen Niels Bohr Institute for Astronomy Physics and Geophysics Astronomical Observatory Copenhagen [email protected] The development of astronomy worldwide begins at the roots: Already from childhood, humans of all nations and civilizations seem to share an innate fascination with the sky. Yet, people in different regions of the world have vastly different possibilities for pursuing this interest. In wealthy, industrialised societies the way is open to a school or higher education in science, possibly leading to a career in astronomy or basic or applied space science for the benefit of the country as well as the individual. In other regions, neither the financial nor the trained human resources are sufficient to offer that avenue to the future of the young generation, or those intellectual resources to the development of their country. This book addresses ways and means by which these obstacles can be, if not fully overcome, then at least significantly reduced.

The papers in this volume present theoretical aspects and applications of artificial neural networks and genetic algorithms. Also included are papers on fuzzy logic, soft computing, and artificial intelligence. Fundamental issues are addressed such as the nonlinear approximation capabilities of neural networks and formal methods of data representation with topological properties. New elements in genetic algorithms are presented, for example, crossover methods and gene representation. Papers on applications of neural networks show how successful these methods are in a wide range of fields like meteorological and atmospheric pollution forecasts, furnace control, and system identification. Genetic algorithms are used to solve optimization problems related to shipping and computer vision. Fuzzy-logic-based techniques are applied to sociodynamic models and hybrid neuro-fuzzy models.

There is hardly a science that is without the notion of "system." We have systems in mathematics, formal systems in logic, systems in physics, electrical and mechanical engineering, architectural-, operating-, infonnation-, programming systems in computer science, management-and PJoduction systems in industrial applications, economical-, ecological-, biological systems, and many more. In many of these disciplines formal tools for system specification, construction, verification, have been developed as well as mathematical concepts for system modeling and system simulation. Thus it is quite natural to expect that systems theory as an interdisciplinary and well established science offering general concepts and methods for a wide variety of applications is a subject in its own right in academic education. However, as can be seen from the literature and from the curricula of university studies -at least in Central Europe-, it is subordinated and either seen as part of mathematics with the risk that mathematicians, who may not be familiar with applications, define it in their own way, or it is treated separately within each application field focusing on only those aspects which are thought to be needed in the particular application. This often results in uneconomical re-inventing and re-naming of concepts and methods within one field, while the same concepts and methods are already well introduced and practiced in other fields. The fundamentals on general systems theory were developed several decades ago. We note the pioneering work of M. A. Arbib, R. E. Kalman, G. 1. Klir, M. D.

This book presents a collection of revised refereed papers selected from the contributions to the Fifth International Workshop on Computer Aided Systems Theory, EUROCAST '95, held in Innsbruck, Austria in May 1995.
The 42 full papers contained have been contributed by CAST theoreticians, tool-makers, designers, and appliers and reflect the full spectrum of activities in the area. The papers are organized in sections on systems theory, design environments, complex systems design, and specific applications.

One ofthe most important aspects in research fields where mathematics is "applied is the construction of a formal model of a real system. As for structural relations, graphs have turned out to provide the most appropriate tool for setting up the mathematical model. This is certainly one of the reasons for the rapid expansion in graph theory during the last decades. Furthermore, in recent years it also became clear that the two disciplines of graph theory and computer science have very much in common, and that each one has been capable of assisting significantly in the development of the other. On one hand, graph theorists have found that many of their problems can be solved by the use of com puting techniques, and on the other hand, computer scientists have realized that many of their concepts, with which they have to deal, may be conveniently expressed in the lan guage of graph theory, and that standard results in graph theory are often very relevant to the solution of problems concerning them. As a consequence, a tremendous number of publications has appeared, dealing with graphtheoretical problems from a computational point of view or treating computational problems using graph theoretical concepts."