|
|||||
|
|
| 混合递阶遗传径向基网络及其在副热带高压预报中的应用 | |
| 引用本文： | 刘科峰,张韧,洪梅,余丹丹,王彦磊,高扬.混合递阶遗传径向基网络及其在副热带高压预报中的应用[J].热带气象学报,2008,24(5):507-511. |
| 作者姓名： | 刘科峰 张韧 洪梅 余丹丹 王彦磊 高扬 |
| 作者单位： | 1. 解放军理工大学气象学院海洋与空间环境系,江苏,南京,211101;中国气象局广州热带海洋气象研究所,广东,广州,510080 2. 解放军理工大学气象学院海洋与空间环境系,江苏,南京,211101 3. 95140部队气象台,广东,广州,516259 |
| 基金项目： | 国家自然科学基金，热带海洋气象科研项目 |
| 摘 要： | 采用遗传算法与径向基网络结合的方法建立了副热带高压特征指数的预报优化模型.针对径向基网络结构和初始参数难以客观确定的不足,引入混合递阶遗传算法同时优化网络结构和参数.该优化方法结合了递阶遗传算法和最小二乘法的优点,具有较高的学习效率.将混合递阶遗传径向基网络用于副高数值预报产品的预报试验和效果比较,结果表明:混合递阶遗传算法优化的径向基网络模型具有较好的收敛效果和泛化能力,对副高指数的预报效果有较明显的改进和提高. |
| 关 键 词： | 混合递阶遗传算法 径向基神经网络 副热带高压 |
| 收稿时间： | 2007/3/12 0:00:00 |
| 修稿时间： | 2007/8/22 0:00:00 |
RBFNN BASED ON HYBRID HIERARCHY GENETIC ALGORITHM AND ITS APPLICATION IN SUBTROPICAL HIGH FORECAST |
|
| LIU Ke-feng,ZHANG Ren,HONG Mei,YU Dan-dan,WANG Yan-lei and GAO Yang.RBFNN BASED ON HYBRID HIERARCHY GENETIC ALGORITHM AND ITS APPLICATION IN SUBTROPICAL HIGH FORECAST[J].Journal of Tropical Meteorology,2008,24(5):507-511. | |
| Authors: | LIU Ke-feng ZHANG Ren HONG Mei YU Dan-dan WANG Yan-lei and GAO Yang |
| Institution: | Institute of Meteorology, PLA University of Science and Technology, Nanjing 211101, China;Guangzhou Insitute of Tropical and Marine Meteorology, CMA, Guangzhou 510080, China;Institute of Meteorology, PLA University of Science and Technology, Nanjing 211101, China;Guangzhou Insitute of Tropical and Marine Meteorology, CMA, Guangzhou 510080, China;Institute of Meteorology, PLA University of Science and Technology, Nanjing 211101, China;Institute of Meteorology, PLA University of Science and Technology, Nanjing 211101, China;Institute of Meteorology, PLA University of Science and Technology, Nanjing 211101, China;95140 Troops Weather Stations, Guangzhou 516259, China |
| Abstract: | A forecasting model of subtropical high was constructed with RBF neural network. But,because of the difficulty of objective determination of the structure and initial parameter of RBF neural network,the hybrid hierarchy genetic algorithm was introduced to optimize network configuration and parameters. The optimization method,which combines the advantages of genetic algorithm and the least squares method,has high learning efficiency. Then,the model was used to forecast the subtropical high. Last,output results from the model were compared with the results of the T106 and RBF network. The results show that the model with RBF neural network based on hybrid hierarchy genetic algorithm improves the subtropical high forecast results from T106 by large margin. |
| Keywords: | hybrid hierarchy genetic algorithm RBF neural network subtropical high |
| 本文献已被 维普 万方数据 等数据库收录！ | |
| 点击此处可从《热带气象学报》浏览原始摘要信息 | |
| 点击此处可从《热带气象学报》下载免费的PDF全文 | |