改进的笛卡尔遗传编程组合电路演化设计

The Improved Cartesian Genetic Programming for Evolution of Combinational Circuit

利用笛卡尔遗传编程进行电路演化时，存在收敛速度慢、收敛时间波动较大等问题．对笛卡尔遗传编程中的可编程单元模型进行改进，增加与目标函数相关的逻辑运算，去除无关的逻辑运算，从而提高演化算法命中目标的概率．利用改进的笛卡尔遗传编程方法分别对电机换相电路和乘法器等组合电路进行演化设计．结果表明，改进后的方法明显缩短了电路演化生成的时间，且收敛时间波动较小．

There are problems existing in evolution of combinational circuit with Cartesian genetic programming. For example, the time consumed by system is very long and not stable. In this pa- per, some improvements are proposed to optimize the structure of programmable elements. We add the logic that have close relation with the target function and remove the logic which has no use. By this way,the probability of achieving the target is enhanced. Then,the commutator of mo- tor and multiplier are evolved with the improved model. The experiment results show that the im- proved reconfigurable circuit model designed by the optimized method could increase the speed of circuit evolving and decrease the fluctuation of evolving time.