Practical Knowledge-Based Systems in Conceptual Design

ISBN-13: 9781447120445 / Angielski / Miękka / 2011 / 243 str.

John C. Miles; Carolynne J. Moore

Opis

Conceptual Design is one of the few areas of Engineering Design where computers have yet to make an impact. With the development of Knowledge Based Systems it is now possible to rectify this situation. This publication deals with the use of Knowledge Based Systems (KBS) as tools for conceptual design. Included are neglected aspects such as evaluation and user needs. Practical Knowledge Based Systems in Conceptual Design is based on the authors' experience of developing KBS for use in civil engineering, an area of industrial application which is recognised as being one of great potential. The methodology has been tried and tested by designers. Examples of systems which have been developed to solve specific design problems are included.

Spis treści

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.