| This dissertation describes a method to store design solutions for a computer-based design environment that supports the early phases of building design. The proposed building design representation, used to store design solutions, consists of two levels of abstraction models defined on top of the object-oriented data model. The first level is a generic information model, called the Building Entity and Technology (BENT) model, that stores design data as they are generated during conceptual design, supports case-based reasoning, and shares data among all design participants. This model represents each building entity as a generic container that encompasses its geometry, taxonomy, design knowledge, composition, relationships, and properties. The second level is a conceptual model that defines the types of objects, relationships, and data needed to fully represent the information in a given design domain. The conceptual model specifies the semantics of the design information for the domain using the syntax defined in the information model. A sample conceptual model has been defined for the structural system of buildings.; The development of case-based reasoning capabilities first involved the development of a faceted classification scheme to define the controlled vocabulary from which indexes are obtained. The classification is extensible and designers have the freedom to complement the vocabulary with their own idiosyncrasies. Indexing is performed automatically as building entities are designed through the selection and application of technologies. Hence, a design is already indexed when it is added to the case library. A case library was implemented in an object-oriented database management system for accumulating cases and for providing efficient query facilities. Three different methods to create a query based on the current problem statement were developed to find the most useful cases stored in the case library. The selected case is adjusted to the current problem in order to provide a complete solution with an adaptation method, called derivational replay, that re-enacts the set of reasoning steps used to produce the retrieved case in the new situation. |
