基于遗传神经网络的地震综合预测系统研究
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
提出利用遗传神经网络技术对地震预测建模,其基本思想是利用遗传算法具有全局搜索能力、不易陷入局部极小点的特点优化神经网络的权重和阈值,利用优化的神经网络挖掘地震前兆异常属性和震级之间的潜在的关系,从而实现地震预测.并选取了《中国震例》中的97个样本数据进行实验,震例验证的结果表明,遗传神经网络算法用于大地震综合预测效果较理想.
A model of earthquake prediction based on genetic-neural network is presented.The main(idea) is that the genetic algorithm can carry through whole-space search and it will not get into local extremum.So the genetic algorithm is able to optimize the proportion and the numeral of the BP algorithm which can search the area that includes whole-space minimum.Hence,genetic-neural network can explore the underlying relation between the abnormity of earthquake and the level of earthquake.The paper utilizes 97 samples which are selected from THE ERRTHQUAKE EXAMPLES OF CHINA to test the model.The results show that the genetic-neural network algorithm is more efficient in the comprehensive earthquake prediction.
引文
[1]梅荣荣,冯德益,张国民.中国地震预报概论[M].北京:地震出版社,1993.
    [2]刘豹.模糊工程决策与决策支持系统[J].管理科学学报,1995,5(3):1-5.
    [3]Schaffer J D,W h itley D.Proceed ings of theW orkshop on Comb inat ions ofGenetic A lgorithm sand NeuralNetworks 1992[M].Los A lam itos,CA:The IEEE Computer Soc iety Press,1992.
    [4]张肇诚.中国震例(1981-1985)[M].北京:地震出版社,1986.
    [5]陈棋福,郑大年,车时.中国震例(1992-1994)[M].北京:地震出版社,1995.
    [6]陈棋福,郑大年,车时.中国震例(1995-1996)[M].北京:地震出版社,1996.
    [7]陈棋福,郑大年,车时.中国震例(1997-1999)[M].北京:地震出版社,2000.
    [8]高玮,郑颖人.岩体参数的进化反演[J].水利学报.2000,30(8):1-5.
    [9]Zhangtong,zhang Jiayu.F loat-Encod ing bybrid genetic algorithm s and its app licationto identification of fuzzy system s[EB/OL].http://166.111.82.3/monograph ictech/suppot/simu l/emu lative5-4.htm l.2004-12-01.
    [10]王小平,曹立明.遗传算法—理论、应用与软件实现[M].西安:西安交通大学出版社,1998.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心