在内部演化硬件中实现自适应变异参数控制

为了在演化过程中优化演化算法性能和避免花费大量时间在演化算法的参数设定中,设计了一种新颖的基于硬件实现的自适应变异比率控制方法.为了实现自适应特性,变异比率控制参数也被编码到染色体中作为附加的基因经历演化操作.本方法的有效性将通过和传统的采用固定变异比率的演化算法在演化4-bit偶校验函数(even-parity function),2-bit乘法器和3-bit乘法器的对比实验中进行证明.实验平台建立在一个完全FPGA实现的内部演化硬件上,它的设计思想来源于笛卡尔遗传程序(Cartesian Genetic Programming).在所有实验中,基于自适应变异比率控制的演化算法的性能明显优于传统的采用固定变异比率的遗传算法.

Implementing self-adaptive mutation rate control on intrinsic evolvable hardware

To allow evolutionary algorithm(EA) optimize its performance in the evolution process and free user from the non-trivial task of parameters decision of EA,it is present a novel hardware based self-adaptive mutation rate control scheme.For achieving the property of self-adaptation,the mutation rate control parameters are encoded as additional genes in chromosome and also undergo the evolutionary operations.The strength of the proposed method is demonstrated by a comparison of its performance with traditional fixed mutation rate EA on the evolutions of a 4-bit even-parity function,a 2-bit multiplier and a 3-bit multiplier.The experimental platform is built on a complete FPGA implemented intrinsic evolvable hardware(EHW) which is inspired by Cartesian Genetic Programming(CGP).Within the range of the experiments,the performances of the proposed scheme are remarkably better than the conventional constant mutation rate based EA.