欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对模糊神经网络控制器通常涉及的参数较多,反传播算法难于收敛的问题,提出了一种优化设计正规化模糊神经网络控制器的量子遗传算法。该方法用量子比特构成染色体,用量子旋转门进行染色体更新,用量子非门进行染色体变异,将量子位的概率幅看作两个并列的基因,因此每条染色体包含两条并列的基因链,在染色体数目相同时,可提高获得全局最优解的概率。对控制器参数随机编码建立初始群体,利用量子遗传算法进行参数优化。实验结果表明该方法是有效的。
英文摘要：
A quantum genetic algorithm was proposed to design the parameters of a normalized fuzzy neural network controller. In this method, chromosomes are comprised of quantum bits, and are updated by quantum rotation gate, and are mutated by quantum non-gate. The probability amplitudes of each quantum bit are regarded as two paratactic genes, each chromosome contains two chains of genes, and each chain of genes represents an optimization result, which can accelerate convergence process and increase successful probability for the same number of chromosomes, The parameters of the normalized fuzzy neural network controller were encoded into some individuals, and the initial colony was formed by some random individuals. A global searching was performed by quantum genetic algorithm. The simulation results show the effectiveness of this method.
同期刊论文项目
同项目期刊论文
期刊信息
位置：