基于多目标自适应遗传算法的逻辑电路门级进化方法
Multi-objective Adaptive Genetic Algorithm for Gate-Level Evolution of Logic Circuits
-
摘要: 提出一种改进的遗传算法,通过网表级编码、多目标评估和遗传参数自适应等措施,可依据多个设计目标,以较少的运算量自动生成和优化逻辑电路.在数字乘法器、偶校验器等进化设计实验中,通过比手工设计和同类方法更优的新奇设计结果展示了该方法的有效性和先进性.Abstract: An adaptive genetic algorithm is proposed for automated design of logic circuitits main features are using a netlist-level representation scheme allowing to change interconnection and select alternative function for each gate-level moduleintroducing a simulation based fitness evaluation method and taking an adaptive strategy to assign genetic parameters.By applying this approach to some basic combinational circuitsincluding digital multipliers and even-parity checkerthe evolution process has generated full functional circuits with novel structureless gates and higher operating speed as compared with that designed by hand-design and some similar approaches.
下载: