FSM Encoding for BDD Representations

Reduced Ordered Binary Decision Diagrams (ROBDDs or simply BDDs) are a data structure used to efficiently represent and manipulate logic functions. They were introduced by Bryant (1986). Since then, they have played a major role in many areas of computer aided design, including logic synthesis, simulation, and formal verification. The size of a BDD representing a logic function depends on the ordering of its variables. For some functions, BDD sizes are linear in the number of variables for one ordering while being exponential for another (Bryant, 1992). Many heuristics have been proposed to find good orderings, e.g., the sifting dynamic reordering algorithm (Rudell, 1993). An exact reordering based on lower bounds was proposed in (Drechsler et al., 2000). An output-efficient algorithm was proposed to realize a reordering transformation (Bern et al., 1996). Other techniques to reduce the BDD size include linear transformations for the variables of the represented function (Gunther