At present one of the main obstacles to a broaderapplication of expert systems is the lack of a theory totell us which problem-solving methods areavailable for agiven problem class. Such a theory could lead to significantprogress in the following central aims of the expert systemtechnique:- Evaluating the technical feasibility of expert systemprojects: This depends on whether there is a suitableproblem-solving method, and if possible a correspondingtool, for the given problem class.- Simplifying knowledge acquisition and maintenance: Theproblem-solving methods provide direct assistance asinterpretation models in knowledge acquisition. Also, theymake possible the development of problem-specific expertsystem tools with graphical knowledge acquisitioncomponents, which can be used even by experts withoutprogramming experience.- Making use of expert systems as a knowledge medium: Thestructured knowledge in expert systems can be used not onlyfor problem solving but also for knowledge communication andtutorial purposes.With such a theory in mind, this book provides a systematicintroduction to expert systems. It describes the basicknowledge representations and the present situation withregard tothe identification, realization, and integrationof problem-solving methods for the main problem classes ofexpert systems: classification (diagnostics), construction,and simulation.|At present one of the main obstacles to a broaderapplication of expert systems is the lack of a theory totell us which problem-solving methods areavailable for agiven problem class. Such a theory could lead to significantprogress in the following central aims of the expert systemtechnique:- Evaluating the technical feasibility of expert systemprojects: This depends on whether there is a suitableproblem-solving method, and if possible a correspondingtool, for the given problem class.- Simplifying knowledge acquisition and maintenance: Theproblem-solving methods provide direct assistance asinterpretation models in knowledge acquisition. Also, theymake possible the development of problem-specific expertsystem tools with graphical knowledge acquisitioncomponents, which can be used even by experts withoutprogramming experience.- Making use of expert systems as a knowledge medium: Thestructured knowledge in expert systems can be used not onlyfor problem solving but also for knowledge communication andtutorial purposes.With such a theory in mind, this book provides a systematicintroduction to expert systems. It describes the basicknowledge representations and the present situation withregard tothe identification, realization, and integrationof problem-solving methods for the main problem classes ofexpert systems: classification (diagnostics), construction,and simulation.|This book provides a systematic introduction to expert systems. It describes the basic knowledge representations and the present situation with regard to the identification, realization, and integration of problem-solving methods for the main problem classes of expert systems.
I Introduction.- 1 Characterization and History of Expert Systems.- 1.1 Characterization.- 1.2 History.- 2 Programming Languages and Expert System Tools.- 2.1 Knowledge Acquisition Problems.- 2.2 Knowledge Acquisition with Specialized Programming Environments.- 2.3 Architecture of Expert Systems.- 3 Use and Usability of Expert Systems.- 3.1 Potential Benefits and Modes of Use.- 3.2 Expert Systems as Knowledge Media.- 3.3 Criteria for Expert System Domains.- 3.4 Summary.- II Basic Techniques of Knowledge Representation.- 4 Logic.- 4.1 Predicate Logic.- 4.2 Properties of Logic Calculi.- 4.3 PROLOG.- 4.4 Summary.- 5 Rules.- 5.1 Forward Chaining.- 5.2 Backward Chaining.- 5.3 Complexity of the Precondition.- 5.4 Rule Structuring.- 5.5 Summary.- 6 Objects/Frames.- 6.1 Example: FRL.- 6.2 Active Objects.- 6.3 Cognitive Meaning of Frames.- 6.4 Automatic Classification in KL-ONE Languages.- 6.5 Frames and Problem-Solving Types.- 6.6 Summary.- 7 Constraints.- 7.1 Types of Constraints and Propagation Algorithms.- 7.2 Solution of a Non-Trivial Constraint Problem.- 7.3 Discussion of the Solution.- 7.4 Formal Characterization of Simple Constraint Systems.- 7.5 Summary.- 8 Probabilistic Reasoning.- 8.1 Bayes´ Theorem.- 8.2 The Dempster-Shafer Theory.- 8.3 The INTERNIST Model.- 8.4 The MYCIN Model.- 8.5 The MED1 Model.- 8.6 Summary.- 9 Non-Monotonic Reasoning.- 9.1 JTMS with Direct Justifications.- 9.2 ATMS with Basic Assumptions as Justifications.- 9.3 Summary.- 10 Temporal Reasoning.- 10.1 Exact Time Relations: VM and MED2.- 10.2 Inexact Quantitative Time Relations: TMM of McDermott.- 10.3 Qualitative Relations: the Time Calculus of Allen.- 10.4 Summary.- III Problem Classes and Problem-Solving Methods.- 11 Previous Approaches to Problem Classification.- 11.1 The Approach of Stefik et al. and Hayes-Roth et al.- 11.2 The Approach of Clancey.- 11.3 The Approach of Breuker et al.- 11.4 The Approach of Chandrasekaran.- 11.5 The Approach of Harmon.- 11.6 The Approach of McDermott.- 11.7 Discussion.- 12 Principles of Problem-Solving Methods.- 12.1 Survey of Problem-Solving Methods.- 12.2 Restriction of Use of Variables.- 12.3 Structure of Knowledge Representation.- 12.4 Structure of Knowledge Manipulation.- 12.5 Structure of Knowledge Acquisition.- Classification.- 13 Survey of the Problem-Solving Type Classification.- 13.1 Domains.- 13.2 Problem Types.- 13.3 Analysis of Problem Characteristics.- 13.4 Problem-Solving Methods.- 14 Simple Classification.- 14.1 Decision Trees.- 14.2 Decision Tables.- 14.3 State of the Art.- 15 Heuristic Classification.- 15.1 Knowledge Representation.- 15.2 Knowledge Manipulation.- 15.3 Knowledge Acquisition.- 15.4 State of the Art.- 15.5 Example: CLASSIKA.- 16 Heuristic Classification: Additional Mechanisms.- 16.1 Treatment of Uncertain Data and Solution Classes.- 16.2 Treatment of Subjective Observations.- 16.3 Recognition of False Observations.- 16.4 Treatment of Time-Dependent Data.- 16.5 Belief Revision.- 16.6 Treatment of Parametrized Data and Solutions.- 16.7 Treatment of Multiple Solutions.- 16.8 Combined Recommendations for Multiple Solutions.- 17 Set-Covering Classification.- 17.1 Knowledge Representation.- 17.2 Knowledge Manipulation.- 17.3 Knowledge Acquisition.- 17.4 State of the Art.- 17.5 Example: FEMO.- 18 Functional Classification.- 18.1 Knowledge Representation.- 18.2 Knowledge Manipulation.- 18.3 Knowledge Acquisition.- 18.4 State of the Art.- 18.5 Example: SIMUL.- 19 Statistical Classification.- 19.1 Knowledge Representation.- 19.2 Knowledge Manipulation.- 19.3 Knowledge Acquisition.- 19.4 State of the Art.- 19.5 Variants of Bayes´ Theorem.- 19.6 Example: FAKTA.- 20 Case-Comparing Classification.- 20.1 Knowledge Representation.- 20.2 Knowledge Manipulation.- 20.3 Knowledge Acquisition.- 20.4 State of the Art.- 20.5 Example: CcC.- V Construction.- 21 Review of the Problem-Solving Type Construction.- 21.1 Domains.- 21.2 Problem Types.- 21.3 Analysis of Problem Characteristics.- 21.4 Problem-Solving Methods.- 22 Skeletal Construction.- 22.1 Knowledge Representation.- 22.2 Knowledge Manipulation.- 22.3 Knowledge Acquisition.- 22.4 State of the Art.- 22.5 Variants of Skeletal Construction: Generate and Test.- 22.6 Heuristic Classification and Skeletal Construction.- 23 Propose and Revise.- 23.1 Knowledge Representation.- 23.2 Knowledge Manipulation.- 23.3 Knowledge Acquisition.- 23.4 State of the Art.- 23.5 Variant for Therapy Planning by Parameter Adjustment.- 23.6 Heuristic Classification and Propose and Revise.- 24 Propose and Exchange.- 24.1 Knowledge Representation.- 24.2 Knowledge Manipulation.- 24.3 Knowledge Acquisition.- 24.4 State of the Art.- 24.5 Example: COKE and REST.- 25 Least-Commitment Strategy.- 25.1 Knowledge Representation.- 25.2 Knowledge Manipulation.- 25.3 Knowledge Acquisition.- 25.4 State of the Art.- 25.5 Example: ExAP.- 26 Model-Based Planning.- 26.1 Non-Hierarchical Planning.- 26.2 Hierarchical Planning.- 26.3 Non-Linear Planning.- 27 Case-Comparing Construction.- 28 Partial Integration of Construction Methods.- 28.1 Knowledge Representation.- 28.2 Knowledge Manipulation.- 28.3 Knowledge Acquisition.- 28.4 State of the Art.- VI Simulation.- 29 Review of the Problem-Solving Type Simulation.- 29.1 Domains.- 29.2 Classification According to Problem Types.- 29.3 Analysis of Problem Characteristics.- 29.4 Problem-Solving Methods.- 30 Single-Phase Simulation.- 31 Numerical Multiple-Phase Simulation.- 31.1 Knowledge Representation.- 31.2 Knowledge Manipulation.- 31.3 Knowledge Acquisition.- 31.4 State of the Art.- 32 Qualitative Multiple-Phase Simulation.- 32.1 Knowledge Representation.- 32.2 Knowledge Manipulation.- 32.3 Knowledge Acquisition.- 32.4 State of the Art.- VII Integration of Problem-Solving Methods.- 33 Basic Ideas for the Integration of Problem-Solving Methods.- 33.1 Domains.- 33.2 Levels of Integration.- 33.3 References to Objects in the Knowledge Representation.- 33.4 Problem Solving.- 33.5 Knowledge Acquisition.- 33.6 Explainability.- 33.7 Data Collection.- 33.8 Programming.- 33.9 Survey of the Integration of Problem-Solving Methods.- 34 Integration of Classification Methods.- 34.1 Strengths and Weaknesses of the Problem-Solving Methods.- 34.2 Knowledge Representation.- 34.3 Knowledge Manipulation.- 34.4 Knowledge Acquisition.- 34.5 Explainability.- 34.6 Data Collection.- 34.7 Programming.- 34.8 State of the Art.- 35 Aspects of the Overall Integration.- 35.1 Survey.- 35.2 Decision-Making in Medicine.- 35.3 Tendering, Configuration and Maintenance of Technical Systems.- 35.4 Job and Production Planning and Quality Control.- Appendix: Survey of Knowledge Representation Formalisms.- References.- System Index.
I Introduction.- 1 Characterization and History of Expert Systems.- 2 Programming Languages and Expert System Tools.- 3 Use and Usability of Expert Systems.- II Basic Techniques of Knowledge Representation.- 4 Logic.- 5 Rules.- 6 Objects/Frames.- 7 Constraints.- 8 Probabilistic Reasoning.- 9 Non-Monotonic Reasoning.- 10 Temporal Reasoning.- III Problem Classes and Problem-Solving Methods.- 11 Previous Approaches to Problem Classification.- 12 Principles of Problem-Solving Methods.- Classification.- 13 Survey of the Problem-Solving Type Classification.- 14 Simple Classification.- 15 Heuristic Classification.- 16 Heuristic Classification: Additional Mechanisms.- 17 Set-Covering Classification.- 18 Functional Classification.- 19 Statistical Classification.- 20 Case-Comparing Classification.- V Construction.- 21 Review of the Problem-Solving Type Construction.- 22 Skeletal Construction.- 23 Propose and Revise.- 24 Propose and Exchange.- 25 Least-Commitment Strategy.- 26 Model-Based Planning.- 27 Case-Comparing Construction.- 28 Partial Integration of Construction Methods.- VI Simulation.- 29 Review of the Problem-Solving Type Simulation.- 30 Single-Phase Simulation.- 31 Numerical Multiple-Phase Simulation.- 32 Qualitative Multiple-Phase Simulation.- VII Integration of Problem-Solving Methods.- 33 Basic Ideas for the Integration of Problem-Solving Methods.- 34 Integration of Classification Methods.- 35 Aspects of the Overall Integration.- Appendix: Survey of Knowledge Representation Formalisms.- References.- System Index.
Inhaltsverzeichnis
I Introduction.- 1 Characterization and History of Expert Systems.- 1.1 Characterization.- 1.2 History.- 2 Programming Languages and Expert System Tools.- 2.1 Knowledge Acquisition Problems.- 2.2 Knowledge Acquisition with Specialized Programming Environments.- 2.3 Architecture of Expert Systems.- 3 Use and Usability of Expert Systems.- 3.1 Potential Benefits and Modes of Use.- 3.2 Expert Systems as Knowledge Media.- 3.3 Criteria for Expert System Domains.- 3.4 Summary.- II Basic Techniques of Knowledge Representation.- 4 Logic.- 4.1 Predicate Logic.- 4.2 Properties of Logic Calculi.- 4.3 PROLOG.- 4.4 Summary.- 5 Rules.- 5.1 Forward Chaining.- 5.2 Backward Chaining.- 5.3 Complexity of the Precondition.- 5.4 Rule Structuring.- 5.5 Summary.- 6 Objects/Frames.- 6.1 Example: FRL.- 6.2 Active Objects.- 6.3 Cognitive Meaning of Frames.- 6.4 Automatic Classification in KL-ONE Languages.- 6.5 Frames and Problem-Solving Types.- 6.6 Summary.- 7 Constraints.- 7.1 Types of Constraints and Propagation Algorithms.- 7.2 Solution of a Non-Trivial Constraint Problem.- 7.3 Discussion of the Solution.- 7.4 Formal Characterization of Simple Constraint Systems.- 7.5 Summary.- 8 Probabilistic Reasoning.- 8.1 Bayes' Theorem.- 8.2 The Dempster-Shafer Theory.- 8.3 The INTERNIST Model.- 8.4 The MYCIN Model.- 8.5 The MED1 Model.- 8.6 Summary.- 9 Non-Monotonic Reasoning.- 9.1 JTMS with Direct Justifications.- 9.2 ATMS with Basic Assumptions as Justifications.- 9.3 Summary.- 10 Temporal Reasoning.- 10.1 Exact Time Relations: VM and MED2.- 10.2 Inexact Quantitative Time Relations: TMM of McDermott.- 10.3 Qualitative Relations: the Time Calculus of Allen.- 10.4 Summary.- III Problem Classes and Problem-Solving Methods.- 11 Previous Approaches to Problem Classification.- 11.1 The Approach of Stefik et al. and Hayes-Roth et al.- 11.2 The Approach of Clancey.- 11.3 The Approach of Breuker et al.- 11.4 The Approach of Chandrasekaran.- 11.5 The Approach of Harmon.- 11.6 The Approach of McDermott.- 11.7 Discussion.- 12 Principles of Problem-Solving Methods.- 12.1 Survey of Problem-Solving Methods.- 12.2 Restriction of Use of Variables.- 12.3 Structure of Knowledge Representation.- 12.4 Structure of Knowledge Manipulation.- 12.5 Structure of Knowledge Acquisition.- Classification.- 13 Survey of the Problem-Solving Type Classification.- 13.1 Domains.- 13.2 Problem Types.- 13.3 Analysis of Problem Characteristics.- 13.4 Problem-Solving Methods.- 14 Simple Classification.- 14.1 Decision Trees.- 14.2 Decision Tables.- 14.3 State of the Art.- 15 Heuristic Classification.- 15.1 Knowledge Representation.- 15.2 Knowledge Manipulation.- 15.3 Knowledge Acquisition.- 15.4 State of the Art.- 15.5 Example: CLASSIKA.- 16 Heuristic Classification: Additional Mechanisms.- 16.1 Treatment of Uncertain Data and Solution Classes.- 16.2 Treatment of Subjective Observations.- 16.3 Recognition of False Observations.- 16.4 Treatment of Time-Dependent Data.- 16.5 Belief Revision.- 16.6 Treatment of Parametrized Data and Solutions.- 16.7 Treatment of Multiple Solutions.- 16.8 Combined Recommendations for Multiple Solutions.- 17 Set-Covering Classification.- 17.1 Knowledge Representation.- 17.2 Knowledge Manipulation.- 17.3 Knowledge Acquisition.- 17.4 State of the Art.- 17.5 Example: FEMO.- 18 Functional Classification.- 18.1 Knowledge Representation.- 18.2 Knowledge Manipulation.- 18.3 Knowledge Acquisition.- 18.4 State of the Art.- 18.5 Example: SIMUL.- 19 Statistical Classification.- 19.1 Knowledge Representation.- 19.2 Knowledge Manipulation.- 19.3 Knowledge Acquisition.- 19.4 State of the Art.- 19.5 Variants of Bayes' Theorem.- 19.6 Example: FAKTA.- 20 Case-Comparing Classification.- 20.1 Knowledge Representation.- 20.2 Knowledge Manipulation.- 20.3 Knowledge Acquisition.- 20.4 State of the Art.- 20.5 Example: CcC.- V Construction.- 21 Review of the Problem-Solving Type Construction.- 21.1 Domains.- 21.2 Problem Types.- 21.3 Analysis of Problem Characteristics.- 21.4 Problem-Solving Methods.- 22 Skeletal Construction.- 22.1 Knowledge Representation.- 22.2 Knowledge Manipulation.- 22.3 Knowledge Acquisition.- 22.4 State of the Art.- 22.5 Variants of Skeletal Construction: Generate and Test.- 22.6 Heuristic Classification and Skeletal Construction.- 23 Propose and Revise.- 23.1 Knowledge Representation.- 23.2 Knowledge Manipulation.- 23.3 Knowledge Acquisition.- 23.4 State of the Art.- 23.5 Variant for Therapy Planning by Parameter Adjustment.- 23.6 Heuristic Classification and Propose and Revise.- 24 Propose and Exchange.- 24.1 Knowledge Representation.- 24.2 Knowledge Manipulation.- 24.3 Knowledge Acquisition.- 24.4 State of the Art.- 24.5 Example: COKE and REST.- 25 Least-Commitment Strategy.- 25.1 Knowledge Representation.- 25.2 Knowledge Manipulation.- 25.3 Knowledge Acquisition.- 25.4 State of the Art.- 25.5 Example: ExAP.- 26 Model-Based Planning.- 26.1 Non-Hierarchical Planning.- 26.2 Hierarchical Planning.- 26.3 Non-Linear Planning.- 27 Case-Comparing Construction.- 28 Partial Integration of Construction Methods.- 28.1 Knowledge Representation.- 28.2 Knowledge Manipulation.- 28.3 Knowledge Acquisition.- 28.4 State of the Art.- VI Simulation.- 29 Review of the Problem-Solving Type Simulation.- 29.1 Domains.- 29.2 Classification According to Problem Types.- 29.3 Analysis of Problem Characteristics.- 29.4 Problem-Solving Methods.- 30 Single-Phase Simulation.- 31 Numerical Multiple-Phase Simulation.- 31.1 Knowledge Representation.- 31.2 Knowledge Manipulation.- 31.3 Knowledge Acquisition.- 31.4 State of the Art.- 32 Qualitative Multiple-Phase Simulation.- 32.1 Knowledge Representation.- 32.2 Knowledge Manipulation.- 32.3 Knowledge Acquisition.- 32.4 State of the Art.- VII Integration of Problem-Solving Methods.- 33 Basic Ideas for the Integration of Problem-Solving Methods.- 33.1 Domains.- 33.2 Levels of Integration.- 33.3 References to Objects in the Knowledge Representation.- 33.4 Problem Solving.- 33.5 Knowledge Acquisition.- 33.6 Explainability.- 33.7 Data Collection.- 33.8 Programming.- 33.9 Survey of the Integration of Problem-Solving Methods.- 34 Integration of Classification Methods.- 34.1 Strengths and Weaknesses of the Problem-Solving Methods.- 34.2 Knowledge Representation.- 34.3 Knowledge Manipulation.- 34.4 Knowledge Acquisition.- 34.5 Explainability.- 34.6 Data Collection.- 34.7 Programming.- 34.8 State of the Art.- 35 Aspects of the Overall Integration.- 35.1 Survey.- 35.2 Decision-Making in Medicine.- 35.3 Tendering, Configuration and Maintenance of Technical Systems.- 35.4 Job and Production Planning and Quality Control.- Appendix: Survey of Knowledge Representation Formalisms.- References.- System Index.
Klappentext
At present one of the main obstacles to a broadernapplication of expert systems is the lack of a theory tontell us which problem-solving methods areavailable for angiven problem class. Such a theory could lead to significantnprogress in the following central aims of the expert systemntechnique:n- Evaluating the technical feasibility of expert systemnprojects: This depends on whether there is a suitablenproblem-solving method, and if possible a correspondingntool, for the given problem class.n- Simplifying knowledge acquisition and maintenance: Thenproblem-solving methods provide direct assistance asninterpretation models in knowledge acquisition. Also, theynmake possible the development of problem-specific expertnsystem tools with graphical knowledge acquisitionncomponents, which can be used even by experts withoutnprogramming experience.n- Making use of expert systems as a knowledge medium: Thenstructured knowledge in expert systems can be used not onlynfor problem solving but also for knowledge communication andntutorial purposes.nWith such a theory in mind, this book provides a systematicnintroduction to expert systems. It describes the basicnknowledge representations and the present situation withnregard tothe identification, realization, and integrationnof problem-solving methods for the main problem classes ofnexpert systems: classification (diagnostics), construction,nand simulation.
This book provides a systematic introduction to expert systems. It describes the basic knowledge representations and the present situation with regard to the identification, realization, and integration of problem-solving methods for the main problem classes of expert systems.
