Effect of BDD Optimization on Synthesis of Reversible and Quantum Logic

Synthesis of reversible and quantum logic has become an intensely studied topic in the last years. However, most synthesis methods are limited, since they rely on a truth table representation of the function to be synthesized. BDD-based synthesis offers an alternative. Here, reversible or quantum circuits are derived from a function given as Binary Decision Diagram (BDD) by substituting all nodes of the BDD with a cascade of Toffoli or elementary quantum gates, respectively. As a result, the application of the approach is not limited by the truth table of the function but by the (quite more efficient) BDD representation. Furthermore, many optimization techniques for BDDs exist which can be exploited. In this work, we evaluate the effect of three optimization methods for BDDs (namely shared nodes, complement edges, and advanced orderings) on the resulting reversible and quantum circuits. We describe in detail the adjustments, which have to be done to support these optimizations for synthesis, and discuss possible improvements and drawbacks. In a case study, the effects are experimentally evaluated. The results showed, that applying these optimization techniques leads to significant smaller circuits (with respect to number of gates and lines) in most of the cases.