Reference Point-based Evolutionary Multi-objective Optimization for Reversible Logic Circuit Synthesis

In this paper, Reversible logic circuit synthesis is formulated as a quantum cost-minimization problem with equality constraint. A new reference-point based evolutionary multi-objective method R-EMO-RLC is specially designed to attack the equality constraint. First, the reference point is determined dynamically according the distribution of solutions. Then, a new crowding comparative operator is fabricated to adapt the uncertainty of constraint violation and objective value aroused by variable length encoding. Experimental results show that R-EMO-RLC can increase the feasible ratio and obtain savings in quantum cost for some benchmarks from recent publications comparing with previously known circuits.