Designing Combinational Circuits Using a Multi-objective Cartesian Genetic Programming with Adaptive Population Size

摘要

This paper proposes a multiobjective Cartesian Genetic Programming with an adaptive population size to design approximate digital circuits via evolutionary algorithms, analyzing the trade-off between the most often used objectives: error, area, power dissipation, and delay. Combinational digital circuits such as adders, multipliers, and arithmetic logic units (ALUs) with up to 16 inputs and 370 logic gates are considered in the computational experiments. The proposed method was able to produce approximate circuits with good operational characteristics when compared with other methods from the literature.