DocumentCode
276563
Title
A Steiner tree construction for VLSI routing
Author
Kahng, Andrew B.
Author_Institution
Dept. of Comput. Sci., California Univ., Los Angeles, CA, USA
Volume
i
fYear
1991
fDate
8-14 Jul 1991
Firstpage
133
Abstract
Proposes a parallel, analog approach for the construction of a minimum rectilinear Steiner tree (MRST), which intuitively shrinks a bubble around the pins of the signal net until a Steiner tree topology is induced. This method maps well to parallel neural-style architectures, as well as to fairly generic two-dimensional array topologies. The author describes the basic approach along with extensive preliminary simulation results which show better performance than existing MRST approaches (i.e., almost 10% average improvement over minimum spanning tree cost). The result is of practical interest for VLSI CAD applications, and is an instance where an analog heuristic for an NP-complete problem outperforms existing combinatorial methods, both in time complexity and average-case performance
Keywords
VLSI; circuit layout CAD; computational complexity; neural nets; trees (mathematics); CAD applications; NP-complete problem; VLSI routing; analog heuristic; bubble shrinking; minimum rectilinear Steiner tree; minimum spanning tree cost; neural-style architectures; parallel analogue approach; performance; signal net pins; simulation; time complexity; tree topology; two-dimensional array topologies; Circuit synthesis; Costs; Design automation; NP-complete problem; Pins; Routing; Steiner trees; Topology; Very large scale integration; Wire;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on
Conference_Location
Seattle, WA
Print_ISBN
0-7803-0164-1
Type
conf
DOI
10.1109/IJCNN.1991.155164
Filename
155164
Link To Document