The nationwide research project Deduktion', funded by the Deutsche Forschungsgemeinschaft (DFG)' for a period of six years, brought together almost all research groups within Germany engaged in the field of automated reasoning. Intensive cooperation and exchange of ideas led to considerable progress both in the theoretical foundations and in the application of deductive knowledge. This three-volume book covers these original contributions moulded into the state of the art of automated deduction. The three volumes are intended to document and advance a development in the field of automated deduction that can now be observed all over the world. Rather than restricting the interest to purely academic research, the focus now is on the investigation of problems derived from realistic applications. In fact industrial applications are already pursued on a trial basis. In consequence the emphasis of the volumes is not on the presentation of the theoretical foundations of logical deduction as such, as in a handbook; rather the books present the concepts and methods now available in automated deduction in a form which can be easily accessed by scientists working in applications outside of the field of deduction. This reflects the strong conviction that automated deduction is on the verge of being fully included in the evolution of technology. Volume I focuses on basic research in deduction and on the knowledge on which modern deductive systems are based. Volume II presents techniques of implementation and details about system building. Volume III deals with applications of deductive techniques mainly, but not exclusively, to mathematics and the verification of software. Each chapter was read bytwo referees, one an international expert from abroad and the other a knowledgeable participant in the national project. It has been accepted for inclusion on the basis of these review reports. Audience: Researchers and developers in software engineering, formal methods, certification, verification, validation, specification of complex systems and software, expert systems, natural language processing.

The nationwide research project Deduktion', funded by the Deutsche Forschungsgemeinschaft (DFG)' for a period of six years, brought together almost all research groups within Germany engaged in the field of automated reasoning. Intensive cooperation and exchange of ideas led to considerable progress both in the theoretical foundations and in the application of deductive knowledge. This three-volume book covers these original contributions moulded into the state of the art of automated deduction. The three volumes are intended to document and advance a development in the field of automated deduction that can now be observed all over the world. Rather than restricting the interest to purely academic research, the focus now is on the investigation of problems derived from realistic applications. In fact industrial applications are already pursued on a trial basis. In consequence the emphasis of the volumes is not on the presentation of the theoretical foundations of logical deduction as such, as in a handbook; rather the books present the concepts and methods now available in automated deduction in a form which can be easily accessed by scientists working in applications outside of the field of deduction. This reflects the strong conviction that automated deduction is on the verge of being fully included in the evolution of technology. Volume I focuses on basic research in deduction and on the knowledge on which modern deductive systems are based. Volume II presents techniques of implementation and details about system building. Volume III deals with applications of deductive techniques mainly, but not exclusively, to mathematics and the verification of software. Each chapter was read bytwo referees, one an international expert from abroad and the other a knowledgeable participant in the national project. It has been accepted for inclusion on the basis of these review reports. Audience: Researchers and developers in software engineering, formal methods, certification, verification, validation, specification of complex systems and software, expert systems, natural language processing.

The nationwide research project `Deduktion', funded by the `Deutsche Forschungsgemeinschaft (DFG)' for a period of six years, brought together almost all research groups within Germany engaged in the field of automated reasoning. Intensive cooperation and exchange of ideas led to considerable progress both in the theoretical foundations and in the application of deductive knowledge. This three-volume book covers these original contributions moulded into the state of the art of automated deduction. The three volumes are intended to document and advance a development in the field of automated deduction that can now be observed all over the world. Rather than restricting the interest to purely academic research, the focus now is on the investigation of problems derived from realistic applications. In fact industrial applications are already pursued on a trial basis. In consequence the emphasis of the volumes is not on the presentation of the theoretical foundations of logical deduction as such, as in a handbook; rather the books present the concepts and methods now available in automated deduction in a form which can be easily accessed by scientists working in applications outside of the field of deduction. This reflects the strong conviction that automated deduction is on the verge of being fully included in the evolution of technology. Volume I focuses on basic research in deduction and on the knowledge on which modern deductive systems are based. Volume II presents techniques of implementation and details about system building. Volume III deals with applications of deductive techniques mainly, but not exclusively, to mathematics and the verification of software. Each chapter was read by two referees, one an international expert from abroad and the other a knowledgeable participant in the national project. It has been accepted for inclusion on the basis of these review reports. Audience: Researchers and developers in software engineering, formal methods, certification, verification, validation, specification of complex systems and software, expert systems, natural language processing.

This book is for both developer and decision makers of R/3 implementation teams who need to understand in-depth and practically the benefits, financial risks and technical backgrounds of IDocs and ALE in interface development. It describes the implementation of interfaces in an R/3 roll-out, imporatnt technologies such as RFC, OLE and Workflow and common standards like EDIFACT, ANSI X.12 or XML. A large number of recipes deliver templates as a starting point for own enhancements. It is for everybody who depends on fast and cost-effective solutions for EDI and it also discusses why many EDI projects are ten times as expensive as they could be. Preparing the reader with the essential knowledge to survive the outrageously fast growing world of data communication and ecommerce via internet and intranet, the book shows in a destilled manner how enterprises using R/3 can efficiently implement Electronic Data Interchange (EDI) both with external partner and with inhouse satellite systems. This book in the tradition of IT-cookbooks, where the reader will find quick recipes and reliable information to cover all aspects of SAP Interfacing and quickly became a standard work for the R/3 world.

The nationwide research project Deduktion', funded by the Deutsche Forschungsgemeinschaft (DFG)' for a period of six years, brought together almost all research groups within Germany engaged in the field of automated reasoning. Intensive cooperation and exchange of ideas led to considerable progress both in the theoretical foundations and in the application of deductive knowledge. This three-volume book covers these original contributions moulded into the state of the art of automated deduction. The three volumes are intended to document and advance a development in the field of automated deduction that can now be observed all over the world. Rather than restricting the interest to purely academic research, the focus now is on the investigation of problems derived from realistic applications. In fact industrial applications are already pursued on a trial basis. In consequence the emphasis of the volumes is not on the presentation of the theoretical foundations of logical deduction as such, as in a handbook; rather the books present the concepts and methods now available in automated deduction in a form which can be easily accessed by scientists working in applications outside of the field of deduction. This reflects the strong conviction that automated deduction is on the verge of being fully included in the evolution of technology. Volume I focuses on basic research in deduction and on the knowledge on which modern deductive systems are based. Volume II presents techniques of implementation and details about system building. Volume III deals with applications of deductive techniques mainly, but not exclusively, to mathematics and the verification of software. Each chapter was read by two referees, one an international expert from abroad and the other a knowledgeable participant in the national project. It has been accepted for inclusion on the basis of these review reports. Audience: Researchers and developers in software engineering, formal methods, certification, verification, validation, specification of complex systems and software, expert systems, natural language processing.

1. BASIC CONCEPTS OF INTERACTIVE THEOREM PROVING Interactive Theorem Proving ultimately aims at the construction of powerful reasoning tools that let us (computer scientists) prove things we cannot prove without the tools, and the tools cannot prove without us. Interaction typi cally is needed, for example, to direct and control the reasoning, to speculate or generalize strategic lemmas, and sometimes simply because the conjec ture to be proved does not hold. In software verification, for example, correct versions of specifications and programs typically are obtained only after a number of failed proof attempts and subsequent error corrections. Different interactive theorem provers may actually look quite different: They may support different logics (first-or higher-order, logics of programs, type theory etc.), may be generic or special-purpose tools, or may be tar geted to different applications. Nevertheless, they share common concepts and paradigms (e.g. architectural design, tactics, tactical reasoning etc.). The aim of this chapter is to describe the common concepts, design principles, and basic requirements of interactive theorem provers, and to explore the band width of variations. Having a 'person in the loop', strongly influences the design of the proof tool: proofs must remain comprehensible, - proof rules must be high-level and human-oriented, - persistent proof presentation and visualization becomes very important."

We are invited to deal with mathematical activity in a sys tematic way [ ... ] one does expect and look for pleasant surprises in this requirement of a novel combination of psy chology, logic, mathematics and technology. Hao Wang, 1970, quoted from(Wang, 1970). The field of mathematics has been a key application area for automated theorem proving from the start, in fact the very first automatically found the orem was that the sum of two even numbers is even (Davis, 1983). The field of automated deduction has witnessed considerable progress and in the last decade, automated deduction methods have made their way into many areas of research and product development in computer science. For instance, deduction systems are increasingly used in software and hardware verification to ensure the correctness of computer hardware and computer programs with respect to a given specification. Logic programming, while still falling somewhat short of its expectations, is now widely used, deduc tive databases are well-developed and logic-based description and analysis of hard-and software is commonplace today.

Since both the coments and the structure of the book appeared to be successful, only minor changes were made. In particular, some recent work in ATP has been incorporated so that the book continues to reflect the state of the art in the field. The most significant change is in the quality of the layout including the removal of a number of inaccuracies and typing errors. R. Caferra, E. Eder, F. van der Linden, and J. Muller have caught vanous minor errors. P. Haddawy and S.T. Pope have provided many stilistic improvements of the English text. Last not least, A. Bentrup and W. Fischer have produced the beautiful layout. The extensive work of typesetting was financally supported within ESPRIT pro ject 415. Munchen, September 1986 W. Bibel PREFACE Among the dreams of mankind is the one dealing with the mechanization of human thought. As the world today has become so complex that humans apparently fail to manage it properly with their intellectual gifts, the realization of this dream might be regarded even as something like a necessity. On the other hand, the incredi ble advances in computer technology let it appear as a real possibility."

This volume contains the elaborated and harmonized versions of seven lectures given at the first Advanced Course in Artificial Intelligence, held in Vignieu, France, in July 1985. Most of them were written in tutorial form; the book thus provides an extremely valuable guide to the fundamental aspects of AI. In the first part, Delgrande and Mylopoulos discuss the concept of knowledge and its representation. The second part is devoted to the processing of knowledge. The contribution by Huet shows that both computation and inference or deduction are just different aspects of the same phenomenon. The chapter written by Stickel gives a thorough and knowledgeable introduction to the most important aspects of deduction by some form of resolution. The kind of reasoning that is involved in inductive inference problem solving (or programming) from examples, and in learning, is covered by Biermann. The tutorial by Bibel covers the more important forms of knowledge processing that might play a significant role in common sense reasoning. The third part of the book focuses on logic programming and functional programming. Jorrand presents the language FP2, where term rewriting forms the basis for the semantics of both functional and parallel programming. In the last chapter, Shapiro gives an overview of the current state of concurrent PROLOG.

This volume contains the elaborated and harmonized versions of seven lectures given at the first Advanced Course in Artificial Intelligence, held in Vignieu, France, in July 1985. Most of them were written in tutorial form; the book thus provides an extremely valuable guide to the fundamental aspects of AI. In the first part, Delgrande and Mylopoulos discuss the concept of knowledge and its representation. The second part is devoted to the processing of knowledge. The contribution by Huet shows that both computation and inference or deduction are just different aspects of the same phenomenon. The chapter written by Stickel gives a thorough and knowledgeable introduction to the most important aspects of deduction by some form of resolution. The kind of reasoning that is involved in inductive inference problem solving (or programming) from examples, and in learning, is covered by Biermann. The tutorial by Bibel covers the more important forms of knowledge processing that might play a significant role in common sense reasoning. The third part of the book focuses on logic programming and functional programming. Jorrand presents the language FP2, where term rewriting forms the basis for the semantics of both functional and parallel programming. In the last chapter, Shapiro gives an overview of the current state of concurrent PROLOG.

This book contains a selection of revised papers and state-of-the-art overviews on current trends and future perspectives of fuzzy systems. A major aim is to address theoretical as well as application-oriented issues and to contribute to the foundation of concepts, methods, and tools in this field. The book is written by researchers who attended the workshop "Fuzzy Systems '93 - Management of Uncertain Information" (Braunschweig, Germany, October 21-22, 1993), organized by the German Society of Computer Science (GI), the German Computer Science Academy (DIA), and the University of Braunschweig.Dieses Buch enthalt ausgewahlte und auf neuesten Stand gebrachte Fachaufsatze und "State of the Art"-Ubersichtsartikel in englischer Sprache. Sie geben einen Uberblick uber aktuelle Trends sowie Zukunftsperspektiven der Fuzzy-Systeme. Besonderer Wert wird darauf gelegt, dass das Buch in einem ausgewogenen Verhaltnis von Theorie und Praxis zur Fundierung von Konzepten, Methoden und Werkzeugen beitragt. Hervorgegangen ist das Werk aus einem von der Gesellschaft fur Informatik (GI), der Deutschen Informatik Akademie (DIA) und der TU Braunschweig gemeinsam veranstalteten GI-Workshop "Fuzzy-Systeme '93 - Management unsicherer Informationen" (Braunschweig, 21.-22.10.1993). Die Aufsatze wurden uberarbeitet und um Uberblicksartikel erganzt, geschrieben von H. J. Zimmermann, H. Hellendorn, D. Nauck, C. Freksa, S. Gottwald und K. D. Meyer-Gramann.

The Computational Brain, das aussergewohnliche Buch uber vergleichende Forschung in den Bereichen von menschlichem Gehirn und neuesten Moglichkeiten der Computertechnologie, liegt hiermit erstmals in deutscher Sprache vor. Geschrieben von einem fuhrenden Forscherteam in den USA, ist es eine Fundgrube fur alle, die wissen wollen, was der Stand der Wissenschaft auf diesem Gebiet ist. Die Autoren fuhren die Bereiche der Neuroinformatik und Neurobiologie mit gut ausgesuchten Beispielen und der gebotenen Hintergrundinformation gekonnt zusammen. Das Buch wird somit nicht nur dem Fachwissenschaftler sondern auch dem interdisziplinaren Interesse des Informatikers und des Biologen auf eine hervorragende Weise gerecht.
Ubersetzt wurde das Buch von Prof. Dr. Steffen Holldobler und Dipl.-Biol. Claudia Holldobler, einem Informatiker und einer Biologin.
Rezension in Spektrum der Wissenschaft
nr. 10, S. 122 f. im Oktober 1997
(...) Die 1992 erschienene amerikanische Originalausgabe des vorliegenden Werkes ist so erfolgreich, dass man bereits von einem Klassiker reden kann. (...)
(...) ....ist das Buch sehr zu empfehlen. In Verbindung von Neurobiologie und Neuroinformatik
konkurrenzlos, vermittelt es einiges von der Faszination theoretischer Hirnforschung, die auch in Deutschland zunehmend mehr Wissenschaftler in ihren Bann schlagt.

Rezension erschienen in: Computer Spektrum 3/1997, S. 2
(...)Das Buch wird somit nicht nur dem Fachwissenschaftler, sondern auch den interdisziplinaren Interesse des Informatikers und des Biologen auf eine hervorragende Weise gerecht(...)"

Das Handbuch deckt alle Facetten des Web Mining ab. Zunachst wird der Prozess des Web Mining ausfuhrlich beschrieben, wobei insbesondere auf den Aspekt des Preprocessing der internetspezifischen Daten eingegangen wird. Besonderer Wert wird auf die zahlreichen Einsatzpotenziale des Web Mining gelegt, wobei grundsatzliche Uberlegungen mit den Ergebnissen bereits realisierter Projekte erganzt werden.
"

Die Essays behandeln privates Erfahrungswissen und wissenschaftliche Erkenntnisse zu den Themenbereichen Korper, Geist, Seele, zwischenmenschliche Beziehungen,
gesellschaftliche, wirtschaftliche und politische Strukturen sowie Wissenschaft, Religion
und Kunst.
"

Das Buch stellt anhand von Praxisfallen dar, wie mit Hilfe der Verfahren des Data Mining und der Business Intelligence Verhaltensmuster und Wissen in grossen Datenbestanden entdeckt werden konnen. Es geht beispielsweise um Kundensegmentierung, Bonitatsprufung oder Werbetragerplanung in Branchen wie Versandhandel, Versicherung, Einzelhandel oder Telekommunikation.

Das Buch behandelt die wichtigsten Methoden zur Erkennung und Extraktion von "Wissen" aus numerischen und nichtnumerischen Datenbanken in Technik und Wirtschaft. Hierzu gehoren Algorithmen zur Vorverarbeitung, Aufbereitung, Visualisierung und Analyse von Daten. Neben linearen statistischen Methoden werden moderne Verfahren aus den Gebieten Clusteranalyse, Fuzzy-Logik, Neuroinformatik, maschinelles Lernen, Entscheidungsbaume und Agentensysteme vorgestellt.

Sehen ist die Ermittlung von Informationen aus Bildern. Welche Informationsquellen dabei genutzt werden und wie die Auswertung im Einzelnen vorgenommen werden kann, ist Gegenstand dieses einfuhrenden Lehrbuches. Es behandelt die sogenannte Kompetenztheorie des Sehens fur die elementaren Wahrnehmungen, wie Kontrast, Farbe, Tiefe und Bewegung. Als visuell gesteuerte Verhaltensleistungen werden Augenbewegungen und die Navigation behandelt. Technisches Sehen (Computer vision) und die Wahrnehmungsmechanismen des Menschen werden wo immer moglich gemeinsam und vergleichend dargestellt. Die verwendeten mathematischen Verfahren werden im Text eingefuhrt und erlautert; ein Glossar wesentlicher Begriffe erleichtert das Verstandnis."

Alle Prozesse in der Natur enthalten eine oder mehrere ungewisse Komponenten, zeigen Ungewissheiten oder haben einen mehr oder weniger ungewissen Ausgang. Dabei kann man unterscheiden, ob man einen Vorgang -oder einen Teil davon -als ungewiss ansieht, weil man ihn nicht exakt deterministisch erfassen kann (z. B. die Kursentwicklung an einer Wertpapierboerse), ob man ihn als genuin zufallig ansieht (z. B. den radioaktiven Zerfall eines Stoffes) oder ob die Ungewissheit des Vorgangs von seiner Beschreibung mit vagen Begriffen herruhrt. Unsere heutigen sehr kom- plexen sozialen und technischen Strukturen sind ohne den Einsatz von Verfahren zur Behandlung ungewisser Effekte nicht mehr vorstellbar, wenn man z. B. nur an Lebens-und Krankenversicherungen einerseits und an die Berechnung der Zu- verlassigkeit technischer Systeme und Prozesse andererseits denkt. Die Entwicklung mathematischer Werkzeuge zur Wahrscheinlichkeitsrechnung und Statistik fuhrte zu der bis in unser Jahrhundert unangefochtenen Stellung der Stochastik als der besten wissenschaftlichen Methode zur Behandlung von Aspekten der Ungewissheit. In der zweiten Halfte des 20. Jahrhunderts etablierte sich dann die Fuzzy Theorie, die Lotfi Zadeh in der Arbeit Fuzzy Sets (1965) als Verallgemeinerung der Can- torschen Mengentheorie begrundete, als eine ernstzunehmende Konkurrentin fur die Aufgabe, Ungewissheiten zu modellieren. Die weiteren Entwicklungen brachten eine uber Jahrzehnte gefuhrte Auseinandersetzung zwischen Stochastikern und Vertre- tern der Fuzzy Theorie, aber auch eine uberaus erfolgreiche Anwendung der Theorie in vielen Bereichen der angewandten Wissenschaften und der Industrie.

Kunstliche Neuronale Netze, die Fuzzy Set Theorie und Evolutionare Algorithmen werden als innovative und komplementare Problemlosungsansatze heute unter dem Begriff Soft Computing oder auch Computational Intelligence zusammengefasst. Das Buch bietet eine kompakte, gut verstandliche Einfuhrung in die Thematik und dokumentiert aktuelle betriebswirtschaftliche Anwendungen und Forschungsprojekte des Soft Computing. Damit erschliesst sich Innovationspotential, das zu Kostenvorteilen und Effizienzsteigerungen fuhren kann. Zahlreiche Abbildungen erganzen den Text. Das Buch kann als Grundlage zur Entwicklung eigener Anwendungen dienen oder als begleitender Text fur Lehrveranstaltungen.

Die Gesellschaft fur Informatik (GI) zeichnet jedes Jahr eine Informatikdisser tation durch einen Preis aus. Die Auswahl dieser Dissertation stutzt sich auf die von den Universitaten und Hochschulen fur diesen Preis vorgeschlagenen Dissertationen. Somit sind die Teilnehmer an dem Auswahlverfahren der GI bereits als "Preistrager" ihrer Hochschule ausgezeichnet. Der Ausschuss der GI, der den Preistrager aus der Reihe der vorgeschlagenen Kandidaten nominiert, veranstaltete in Raumen der Akademie der Wissen schaften und Literatur Mainzein Kolloquium, das den Kandidaten Gelegenheit bot, ihre Resultate im Kreis der Mitbewerber vorzustellen und zu verteidigen. Der Ausschuss war von dem hohen Niveau der eingereichten Arbeiten und der Prasentation sehr positiv beeindruckt. Die Teilnehmer begrussten die Veran staltung des Kolloquiums sehr, nahmen an der Diskussion teil und schatzten die Moglichkeit, mit den Teilnehmern aus anderen Hochschulen ins Gesprach zu kommen. Zu dem Erfolg des Kolloquiums trug auch die grosszugige Gast freundschaft der Akademie bei, der hier dafur auch gedankt sei. Es fiel dem Ausschuss schwer, unter den nach dem Kolloquium in die engere Wahl genommenen Kandidaten den Preistrager zu bestimmen. Die Publikation der hier prasentierten Kurzfassungen gleicht die Ungerechtigkeit der Auswahl eines Kandidaten unter mehreren ebenburtigen Kandidaten etwas aus."

Neuronale Netze sind in den letzten Jahren Gegenstand intensiver Forschungen gewesen. Dieses Buch verbindet die Darstellung neuester Ergebnisse aus dem Bereich der Lernverfahren mit anwendungsbezogenen Aspekten. Es werden methodische Prinzipien der Erstellung von Softwaresystemen, die auf konnektionistischen Verfahren basieren, herausgearbeitet. Fallbeispiele aus unterschiedlichen Anwendungsdomanen zeigen die vielfaltigen Einsatzmoglichkeiten fur Neuronale Netze.

Dieses Buch ist das Standardwerk zu einem neuen Bereich der angewandten Fuzzy-Technologie, der Fuzzy-Clusteranalyse. Diese beinhaltet Verfahren der Mustererkennung zur Gruppierung und Strukturierung von Daten. Dabei werden im Gegensatz zu klassischen Clustering-Techniken die Daten nicht eindeutig zu Klassen zugeordnet, sondern Zugehorigkeitsgrade bestimmt, so dass die Fuzzy-Verfahren robust gegenuber gestorten oder verrauschten Daten sind und fliessende Klassenubergange handhaben konnen.
Dieses Werk gibt eine methodische Einfuhrung in die zahlreichen Fuzzy-Clustering-Algorithmen mit ihren Anwendungen in den Bereichen Datenanalyse, Erzeugung von Regeln fur Fuzzy-Regler, Klassifikations- und Approximationsprobleme sowie eine ausfuhrliche Darstellung des Shell-Clustering zur Erkennung von geometrischen Konturen in Bildern."

Dieses Lehrbuch aus dem KI-Themenfeld richtet sich an Wirtschaftsinformatiker und Informatiker, ferner an Ingenieure und OR-Spezialisten. Es bietet eine umfassende methodisch orientierte Einfuhrung in das Optimieren mit Evolutionaren Algorithmen. Dazu gehoren vor allem Genetische Algorithmen, Evolutionsstrategien, Genetische bzw. Evolutionare Programmierung. Wichtige Ergebnisse der Theorie werden in gut verstandlicher Form wiedergegeben. Zahlreiche Abbildungen und Beispiele sowie Hinweise auf Quellen im Internet und Testdaten erganzen den Text. Das Buch kann als Grundlage zur Entwicklung eigener Anwendungen dienen oder als begleitender Text fur Lehrveranstaltungen."