摘要:
演化硬件作为新的硬件载体,具有自组织、自适应、自修复的能力,是人工智能在高能激光控制方面的一个重要应用。遗传算法是影响硬件演化速度的一个重要因素。针对目前传统遗传算法进化时间长、运算量大的问题,提出了一种改进的猴王遗传算法种群杂交猴王遗传算法。受自然界生物种群杂交优势的启发,种群杂交猴王遗传算法将参与进化的基因序列划分为Nd个独立进化的子种群。每个子种群独立按照猴王遗传算法进化Td代形成原始种群的Nd个亚种群后,交换亚种群的猴王基因重复猴王遗传操作,在亚种群中产生具有杂交优势的后代。分析表明:与猴王遗传算法相比,种群杂交猴王遗传算法可以将每一代基因排序的运算量减小到1/Nd,并且更加利于并行实现。基于MATLAB和Modelsim的仿真分析表明：种群杂交猴王遗传算法具有更快的收敛速度和更优的进化结果。
Abstract:
The evolution hardware as a new hardware carrier, having self-organizing, adaptive, self-repair ability, is an important application of artificial intelligence in the controller of high energy laser. Genetic algorithm is one of the important factors that influence the hardware evolution speed. For the problems of long evolutionary time and large amount of computation of traditional genetic algorithm, an improved genetic algorithmPopulation Hybridization Monkey-King Genetic Algorithm (PHMKGA)was proposed. Inspired by hybrid vigor in biological species, gene sequences in the PHMKGA were divided into Nd independent evolution sub populations while evolving. Each sub population was formed by evolving according to Monkey-King Genetic Algorithm from the original population; the Monkey king genes of sub populations were exchanged to be repeated Monkey King genetic operation; the offsprings of heterosis were produced in sub population. Analysis shows that the PHMKGA(Nd is the number of sub population)could reduce the computation of gene ordering in each generation to 1/Nd comparing to Monkey-King Genetic Algorithm, and is more conducive to the realization of parallel. The simulation analysis based on MATLAB and Modelsim indicates that the PHMKGA results in faster convergence speed and better evolution.
点击查看大图
下载:
