This chapter introduces the three main themes informing this work: knowledge (the subject of investigation), modelling (the chosen approach), and reuse (the pragmatics of the exercise).
This chapter critically analyses current approaches to knowledge modelling, focusing especially on existing modelling frameworks, libraries, and characterizations of problem solving methods. The analysis highlights a number of issues, concerning library organization and the nature of problem solving methods.
A novel approach to knowledge modelling and library organization is presented. Individual problem solving methods are derived from a generic problem solving model, in turn grounded on a generic paradigm and a formal model of a class of applications. The resulting library enjoys a clear theoretical basis and provides robust support for reuse.
The chapter gives an overview of the modelling language used for formalizing library components and for building the various application models described later in the book. The discussion includes an account of the philosophy underlying the language, its main modelling constructs and a comparison with other modelling languages.
The framework used for modelling tasks and problem solving methods and for organizing libraries of problem solving components is formalized as an ontology expressed in the OCML modelling language.
This chapter situates parametric design applications in the context of the wider class of AI design problems, proposes a task ontology for parametric design and compares it to alternative models.
A generic problem solving model for parametric design is described, which consists of a set of generic tasks and an associated ontology. The tasks provide high-level building blocks for constructing problem solving methods for parametric design. The ontology specifies the minimal commitments which need to be obeyed by any problem solving method for parametric design.
This chapter discusses a number of problem solving methods for parametric design, constructed by specializing the generic problem solving model presented in the previous chapter. Thus, it is shown that the proposed model provides both a clear theoretical basis and a useful engineering foundation for developing libraries of problem solving components for parametric design.
This chapter illustrates a number of application models constructed by applying the library of problem solving components described in the previous chapters to three application domains. The aim of this exercise is to validate the various technologies presented earlier in the book, by showing that they provide effective support for KBS development by reuse.
This chapter summarizes the main contributions of this book, outlines recent developments in knowledge modelling research and looks at the broader issues concerning the role of knowledge modelling technology in the emerging cyber-society.
This appendix provides additional information on the OCML language. It describes the informal semantics of the primitive constructors for functional and control terms, the OCML inheritance mechanism, the interpreters and the proof system.
This appendix provides a complete specification of the task-method ontology discussed in chapter 5.
This appendix provides a complete specification of the parametric design task ontology discussed in chapter 6.
This appendix provides a complete specification of the most generic method in the library. This method acts as a kind of 'generic method template', from which all the other methods in the library can be constructed through a method specialization process.