基于BP神经网络模型的城市自然生态型河流景观方案质量评价  

Landscape Plan Quality Evaluation of Urban Natural Ecotype Rivers Based on BP Neural Network Model

作　　者：乔丽芳[1] 张毅川[1] 齐安国[1] 李新铮[1] 姚连芳[1] 

机构地区：[1]河南科技学院园林学院,河南新乡453003

出　　处：《资源开发与市场》2011年第4期314-315,351,共3页Resource Development & Market

基　　金：2009年河南省科技发展计划重大项目(编号:092101310300);2010年河南省科技计划重点项目(编号:102102310246)

摘　　要：城市自然生态型河流的景观设计方案质量评价受多种因素的影响,这些因素之间存在着复杂的非线性关系,有些甚至是随机的、模糊的,利用传统的方法难以表达它们之间的内在关系。研究建立三层次BP神经网络模型,以30份河流景观设计方案为样本,分别根据水质、水量、水空间、植被、经济、设施、交通等7项指标对景观方案的质量进行评价。结果表明,BP神经网络模型具有极强的非线性逼近能力,能真实反映景观质量与影响因素之间的非线性关系,预测结果与实测值之间误差小,相对误差小于5%。该方法操作性强,结果可靠。Landscape plan quality evaluation of urban natural ecotype rivers were influenced by many kinds of factors which had the complex misalignment relations,some even was stochastic and fuzzy,and it was very difficult to express the intrinsic relations between them using the traditional method.This research established three-level BP neural network model,took 30 rivers landscape design plan as the sample,carried on the appraisal separately to the landscape plan′s quality according to the water quality,the water volume,the hydrospace,the vegetation,the economy,the facility,the traffic and so on.The result indicated that the BP neural network model had the great nonlinear approximation ability,could really reflect the misalignment relations between the landscape quality and influencing factor′s,the error was small between the forecasting result and the actual value,the relative error was smaller than 5%,this method was operational strongly and the result was reliable.

关 键 词：自然生态型河流 景观质量评价 BP神经网络 

分 类 号：X32[环境科学与工程—环境工程] D21[政治法律—政治学]