数字电路模块化进化算法

针对大规模复杂电路进化设计的收敛速度和规模瓶颈,在遗传算法的基础上提出了一种模块化进化算法.该算法以节点作为基本单元,采用图表形式的编码方案,其基本思想是将染色体中优秀的基因片段作为有效局部解或优秀子电路封装为模块,进化过程中该模块不再进行进化操作.不仅保护优秀的基因片段而且大大简化了复杂电路的进化设计.并以加法器和乘法器为例进行了模块化进化算法的进化设计实验.结果表明:相对于传统遗传算法,模块化进化算法应用于复杂电路进化设计时,不仅进化设计的速度得到提高,而且大大提高了电路进化设计的成功率.

Modular Evolutionary Algorithm of Digital Circuit

For the convergence speed and scale bottlenecks of evolutionary design of circuits,a modular evolutionary algorithm( MEA) is proposed on the basis of the original genetic algorithm. Considering the particularity of circuit form,the genotype coding in chart form uses the nodes as basic units. The genetic operation of modular is studied and the experiments are designed with adder and multiplier as examples.The MEA can protect good gene fragment by packaging the local solution and subcircuits as module,and simplify the design of complex circuit. Compared with the traditional genetic algorithm,MEA has an obvious advantage and high success rate in complex circuit evolutionary design.