Practical Knowledge-Based Systems in Conceptual Design

1 Knowledge-Based Systems: An Overview.- Aims of the Book.- Computing Systems and Design.- What are KBS?.- Historical Perspective.- Types of KBS.- Successful KBS Applications.- Will KBS Freeze Expertise? - The Deskilling Argument.- Attitudes to "Computer Mistakes".- Will KBS Make Experts or Others Redundant?.- What will be the Benefits of Design KBS?.- 2 Knowledge.- Objectives.- What is Knowledge?.- Types of Knowledge.- Application to Design.- Further Classifications of Knowledge.- Algorithmic Knowledge and Heuristics.- Simple Example.- Examples from Design.- The Basis of Heuristics.- KBS Heuristics.- Classification of Heuristics.- Short-Cut Heuristics.- Heuristics Based on Background Knowledge.- Heuristics Based on Ill-Defined Concepts.- Heuristics Based on Empirical Data.- Heuristics Provided by Others.- Replacement of Heuristics.- An Example.- Deep and Shallow Knowledge.- Meta-Knowledge.- Bias In Human Decision Making.- Consequences of Bias.- Conclusions.- 3 Uncertainty.- Objectives.- Sources of Uncertainty.- Uncertainty in KBS Development.- How do People Deal with Uncertainty?.- The Case for Pessimism.- The Optimistic View.- The Applicability of Representing Uncertainty in KBS.- Design Versus Diagnostic Systems.- How Do your Experts Deal with Uncertainty?.- Is Uncertainty Relevant in Your Domain?.- Does Uncertainty Representation Breed Uncertainty?.- How Subjective is Uncertainty?.- Is it up to the Expert to Reason Uncertainty out of the Domain?.- Uncertainty Management - An Overview.- Representation and Propagation of Uncertain Information.- Numerical Approaches.- Probability.- Bayesian Probabilities.- Certainty Factors.- The Dempster-Shafer Theory of Evidence.- Fuzzy Logic.- Non-Numerical Approaches.- Logic-Based Approaches to Uncertainty Representation.- Non-Formal Representations of Uncertainty.- Conclusions.- 4 Design.- Objectives.- An Overview of Design.- Definitions of Design.- Historical and Sociological Perspectives.- Components of Design.- Classes of Design.- Subsections of Design.- The Effect of Profession and Organisation.- Types of Profession.- Organisational Structures.- Design: The Reality.- Are Designers Aware of How They Design?.- Protocol Analysis.- So How Do Designers Design?.- Memory, Notes and Sketches.- Chunking of Knowledge and Recall.- Summary of "Real Design".- Systematic Design.- Origins of/Motivation for Systematic Design.- Systematic Design: An Example.- Clarification of the Task.- Conceptual Design.- Embodiment and Detail Design.- Systematic Design: A Summary and a Comparison.- Design and KBS.- Positive Arguments.- Negative Arguments.- The Place of Design KBS.- 5 Knowledge Acquisition and Interpretation.- Objectives.- Assessing Your Target Domain.- Identification of Users and their Requirements.- Choosing the Right Experts.- Can the System Developer be the Expert?.- What Makes an Expert "Right"?.- Finding Co-operative Companies.- Finding the Individual Experts.- The Inarticulate Expert.- The Sceptical Expert.- The Expert Who Fears Replacement or Redundancy.- The Non-Committal Expert.- The Out-of-Date Expert.- The Inaccessible Expert.- The Expert Who Doubts His/Her Own Expertise.- The "Fake" Expert.- Single or Multiple Experts?.- Advantages of Using More than One Expert.- Reduced Inaccessibility.- Alternative Source for Explanations.- Two Different Approaches to the Domain Available.- Less Chance of Missing Vital Information.- Elicitation of the Exceptions to the Rule.- Preferences Owing to Familiarity Counterbalanced.- Reduction in Time.- Immediate Verification Procedure.- Disadvantages of Using Multiple Experts.- Possibility of Clashes of Information.- Questioning of Integrity.- Covering the Same Ground Twice.- The Importance of Assessing How Your Experts Think.- The Factors which Help You to Assess the Experts.- Co-operation.- Enthusiasm.- Articulateness.- Bias.- Confidence.- How Comfortable is Your Expert?.- Choosing an Overall Approach to KE.- Is Rapid Prototyping Suitable?.- Interviewing Techniques.- Unstructured Interviewing.- Structured Interviewing.- Alternative Interviewing Techniques.- Pictorial Representations.- Card Sorts.- Repertory Grids/Triples.- Questionnaires.- Case Studies.- Paper Models.- Brainstorming.- Role Playing.- Choosing Interviewing Techniques.- Recording and Interpreting the Knowledge.- Transcript Analysis.- Intermediate Representation Techniques.- Forms of Intermediate Representation.- Feedback Ill.- Conclusions.- 6 Knowledge Representation and Control.- Objectives.- Artificial Intelligence.- How do People Think?.- Different Set of Interests When Developing KBS.- Which Development System?.- Languages.- Symbolic Languages.- Algorithmic Languages.- C.- Shells and Toolkits.- Which Method of KR?.- When to Choose the Form of Knowledge Representation.- Approaches to Knowledge Representation.- Object-Based Approaches.- Semantic Networks.- Frames.- Object Oriented Programming.- Rule-Based Representation.- Predicate and Logic Programming.- Control and Search Mechanisms.- Backward Chaining.- Forward Chaining.- Hybrid Control Structures.- Search Strategies.- Depth First Search.- Breadth First Search.- Other Search Techniques.- Genetic Algorithms.- Constraint Logic Programming.- Rule Selection.- Meta-Rules.- Decision Lattices.- Conclusions.- 7 Evaluation of Practical Systems.- Objectives.- Evaluation, Verification and Validation.- Evaluation.- Validation.- Verification.- Why Bother with Evaluation?.- Why is Evaluation Important for Design KBS?.- The Difficulties with Evaluation.- How these Difficulties can be Overcome.- Evaluation of Other Systems.- Considerations in Evaluation.- Objective Standards.- Bias.- User Orientation.- Efficiency.- Interaction of Knowledge.- Realistic Time Demand on the Evaluators.- An Overall Summary of the Basic Approaches to Evaluation.- Using Case Studies.- Using Experts.- Using a Broad Range of Interviewing Techniques and Reviewers.- The Reviewers.- The Evaluation Process.- When to Evaluate.- Preliminary Evaluation.- If Rapid Prototyping has not been used.- When Rapid Prototyping has been used.- Long-Term Evaluation.- User Interaction.- Aims of the User Interaction.- Acquiring and Assessing the Evaluation Information.- The Diary.- Follow-Up Interviews.- Listing of Comments.- The Analysis of the Information Obtained.- The Importance of Feedback in Evaluation.- Summary and Conclusions.- 8 Examples of the Implementation of KBS for Conceptual Design.- Objectives.- Literature Review of Design KBS.- The Evolution of Design Knowledge-Based Systems.- Some Recent Examples of Design KBS.- Example 1: A Progressive System.- Example 2: The Benefits of Using a Restricted Domain.- Example 3: Systems Using Co-operating Agents.- Example 4: Case-Based Reasoning.- Detailed Applications.- Case Study 1.- Choice of Domain and Aims of the System.- Software and Hardware.- System Features.- An Example of a System Consultation.- Future Developments.- Case Study 2.- Choice of Domain.- Aims of the System.- Sources of Knowledge.- User Interaction.- Style of Knowledge Elicitation.- System Development.- Knowledge Representation.- System Features.- User Interface.- The Help Facility.- An Example of a System Consultation.- Evaluation.- Future Developments.- Case Study 3.- Initial Work.- Setting up the Domain.- System Objectives.- Hardware/Software.- Knowledge Elicitation.- Knowledge Base.- System Features.- The Input.- The Analysis.- The Output.- An Example.- Future Developments.- Conclusions.- 9 Future Developments.- Objectives -.- The Overall Form of Design Systems.- Data Exchange.- Small is Beautiful?.- Product Models.- KBS for Conceptual Design.- Prescriptive or Reactive?.- What Form of Knowledge Base?.- Uncertainty.- Combining Advanced Computing Techniques.- Improving the Acceptance of KBS in Industry.- Evaluation.- Increased Interaction with Industry.- Increased User Involvement and Control.- Conclusions.- References.