Reversible Logic Synthesis-Oriented Multi-Objective Automatic Design Method Based on Evolutionary Design Techniques

Abstract

This paper applies evolutionary design techniques to the reversible logic synthesis, and proposes a reversible logic synthesis-oriented multi-objective automatic design method based on evolutionary design techniques. Firstly, build a gate-level array model of reversible logic circuits (RLC) in order to model the synthesis problems as ones of constrained multi-objective optimizations. Then, encode the candidate RLC to a set of binary evolutionary individuals which are solved by the specialized Pareto-optimal multi-objective evolutionary algorithm. In addition, adopt “pre-bit priority” mechanism to repair infeasible individuals and adopt the rule-based local transformation method to simple the redundant RLC. The results of synthesis experiments demonstrate that the proposed method is feasible and effective, and can automatically synthesize the better RLC.