Modeling concepts for VLSI CAD objects
ACM Transactions on Database Systems (TODS)
Journal of the ACM (JACM)
Efficient Algorithms for Shortest Paths in Sparse Networks
Journal of the ACM (JACM)
Fundamentals of Computer Alori
Fundamentals of Computer Alori
Data structuring and indexing for data base machines
CAW '80 Proceedings of the fifth workshop on Computer architecture for non-numeric processing
New features for a relational database system to support computer aided design
New features for a relational database system to support computer aided design
Hi-index | 0.00 |
Design configurations of objects in engineering design applications, such as VLSI CAD, mechanical CAD and software design systems, are complex compositions of alternatives and historical versions of other objects. Design configurations can be modeled as nodes in a directed acyclic graph (DAG). Relationships between nodes are represented as edges. The design expansion problem, that of expanding a design object in terms of its components, is reduced to the problem of finding all descendants of a given node in the DAG. The converse problem of determining the objects that contain a given object is the same as finding all ancestors of a node. If the nodes of the DAG are stored randomly in secondary storage, then a page containing multiple nodes of the result set (nodes that are ancestors or descendants) may have to be accessed multiple times. We propose a clustering method that guarantees a single access to any such page, and zero access to any page that does not contain a node of the result set. Further, the intermediate storage necessary to prevent duplicate nodes from entering the result set is shown to be small. The sequence of nodes created by the clustering method satisfies certain properties that are continually maintained during the insertion of new nodes into the DAG. Thus, it is possible to dynamically add new design configurations by expanding the DAG through the insertion of new nodes.