| This thesis presents a data management model to support collaborative design environments. Specifically, the proposed model describes a multidisciplinary project in terms of independently evolving designs from the participating disciplines. The model monitors independent design activities by systematically tracking component descriptions in the individual disciplines. Projects are coordinated through asynchronous communication of design changes. There are two salient features of the given model. First, we specify a three-layered closely coupled framework of versions, assemblies, and configurations. In this framework, configurations integrate designs from each of the participating disciplines to describe an overall project. Assemblies can be either total or partial. The designs in each discipline are represented as total assemblies; partial assemblies represent complex entities that can be further aggregated to describe an overall design in a given discipline. Complex entities are in turn formed by aggregating instances of primitive entities in the database. We maintain an evolving description of a primitive entity as a version hierarchy; each version in a given hierarchy contains specific descriptions of instances of that entity.; The second feature introduces a new concept, equivalent operation, which provides an efficient mechanism for managing changes among versions of a primitive entity. Intuitively, an equivalent operation is a single data operation that summarizes the effect of a sequence of changes on an instance description. By applying this concept, we establish a version of a primitive entity as the summary of all changes that have been made to its contents. A version in the data management model is thus a unit of granularity whose consistency can be evaluated. In addition, we develop operators to store, detect and characterize changes among individual versions. The close coupling of the version, assembly and configuration levels allows these computed version changes to be recursively combined to represent changes at various assembly and configuration levels.; Applying these two concepts, the model efficiently supports project coordination through the asynchronous communication of changes among designers, as well as project monitoring through the systematic tracking of evolving project descriptions. We have implemented and tested the model in both an ORACLE relational database for alpha-numeric design data as well as an AUTOCAD environment for 3-D graphical objects. |
