An intelligent design support environment : the application of intelligent knowledge-based systems and advanced HCI techniques to building design
Building design is becoming an increasingly complex process. Technological advances in building materials and construction methods have necessitated the specification of more rigourous regulatory constraints to which the designer must adhere. Although a diverse range of sophisticated computer based design tools, addressing the formal functional requirements of building design, exist to assist the designer in the decision making process, as a result of their sophistication, such tools often require considerable specialist knowledge of the methodologies employed before they can realistically be utilized on a routine basis. As a result a growing interest has developed in intelligent user-interfaces in an attempt to make complex application software more accessible, maintainable and extendible. However, owing to inconsistencies between front-ends, the current trend in user interface management systems tends to propagate the encapsulation of application functionality within a static, esoteric style of dialogue; restricting interaction to the lowest common user level and therefore denying the designer unrestricted access to the embedded methodologies required for creative solution synthesis. By adopting a communications view of the user-interface, this thesis illustrates how a dynamically adaptable user-interface, coupled to a multi-level knowledge based system consisting of surface level models derived from human laws, with deep models of reasoning, employing non-procedural, opportunistic knowledge acquisition mechanisms, may be utilised to accommodate the dynamically varying nature of the design process. The resulting object oriented framework is an intelligent design support system which isolates the user from the low level aspects of CAD tool management; enabling experts from different sub-disciplines to access the functionality of a comprehensive range of design tools in manner suited to their individual conceptual vocabulary, level of expertise, and idiosyncratic design procedures. Although the framework described within this thesis is generally applicable across a range of domains, specific examples of user stereotypes and dialogue templates used to illustrate the principles behind the system are derived from building performance assessment and prediction.