Partitioning by probability condensation
- 1 January 1989
- proceedings article
- Published by Association for Computing Machinery (ACM)
- p. 758-761
- https://doi.org/10.1145/74382.74523
Abstract
A partitioning model is formulated in which components are assigned probabilities of being placed in bins separated by partitions. The expected number of nets crossing partitions is a quadratic function of these probabilities. Minimization of this expected value forces condensation of the probabilities into a “definite” state representing a very good partitioning. The bipartitioning case is treated explicitly.This publication has 0 references indexed in Scilit: