Functional approaches to generating orderings for efficient symbolic representations

Abstract

The authors present a functional approach to generating orderings for representing functions. They develop a cost function which closely mimics the ordered binary decision diagram operations and can be quickly computed. Using the cost as a metric for an ordering, an annealing procedure was used to arrive at good variable orderings. The results obtained by simulated annealing are compared to orderings generated from heuristics that use circuit topology to arrive at a variable ordering.<>