洪水灾情评价的脉冲耦合神经网络模型  

Pulse-coupled Neural Network Model for Evaluation of Flood Disasters

作　　者：杨聪辉 王宝华 付强[2] 谢永刚[3] 

机构地区：[1]广州市水利水电勘测设计研究院,广东广州510640 [2]东北农业大学水利与建筑学院,黑龙江哈尔滨150030 [3]黑龙江大学生产力研究中心经济与工商管理学院,黑龙江哈尔滨150080

出　　处：《灾害学》2010年第3期12-15,共4页Journal of Catastrophology

基　　金：国家自然科学基金项目(30400275)

摘　　要：洪水灾情的评价工作对洪水灾害的分类管理具有重要的指导意义。脉冲耦合神经网络模型应用到洪水灾情评价中是一次新尝试,该方法通过脉冲输出从而调节阈值,并且应用动态阈值来确定洪水灾情的等级。这个方法比传统的BP模型简化了权值的训练,模型更加简便、直观。采用脉冲耦合神经网络模型得出的评价结果同灰色关联法、灰色聚类法、灰色模糊综合法的评价结果进行了比较,表明其用于洪水损失评价具有科学性和实用性。Evaluation of flood disasters is of an important guiding significance to the classified management of flood disasters. As a new attempt,the pulse-coupled neural network model is used in the evaluation of flood disasters. With this method,threshold value is adjusted by output of impulse and grade of a flood disaster is determined by using dynamic threshold. Compared to BP model,this model simplifies the training of the weight. The model is more simple,convenient and intuitive. Comparison of the results of pulse-coupled neural network model with grey association,grey clustering and synthetic grey-fuzzy shows that pulse-coupled neural network model is of scientific and practical value for evaluation of flood disaster losses.

关 键 词：脉冲耦合神经网络模型 洪水灾情评价 BP模型 

分 类 号：P333.2[天文地球—水文科学]