Reversible computing, Genetic algorithms, Evolutionary computation, Electronic digital computers -- Evaluation
We propose an automated synthesis of Reversible logic (RL) circuits using Darwinian and Lamarckian Genetic Algorithms (GA). Our designs are in a form of cascades of generalized gates which generalize factorized Exclusive-Or-Sum-of-Products (ESOP) circuits. GA can be used to explore the problem space of combinational functions and here it is used to evolve reversible logic circuits. We emphasize the role of problem encoding - a well-designed encoding leads to improved results. Our method with well-encoded circuits is compared to standard method on classical benchmarks in GA, and shows good results for synthesis of both random functions and benchmark functions with practical meaning, such as adders.
Lukac, Martin, Mihail Pivtoraiko, Alan Mishchenko, and Marek Perkowski, "Automated synthesis of generalized reversible cascades using genetic algorithms." 5th International Workshop on Boolean Problems. 2002.