沉降预测的变系数灰色GAM模型研究  

STUDY ON VARIABLE COEFFICIENT GAM MODEL OF SETTLEMENT PREDICTION

作　　者：偶昌宝[1] 俞亚南[1] 张仪萍[1] 王士金[1] 

机构地区：[1]浙江大学建筑工程学院,浙江杭州310027

出　　处：《低温建筑技术》2005年第1期76-78,共3页Low Temperature Architecture Technology

基　　金：国家自然科学基金(50278089);浙江省自然科学基金(501163)

摘　　要：灰色理论在路基沉降预测课题中获得了广泛的应用。本文避开一般灰色建模采用的灰色GM(1,1)模型,而根据灰色建模思想,用代数多项式拟合离散数据列,建立了灰色代数曲线模型。将代数多项式的系数设为时间的参数即变系数,并运用拉普拉斯条件极值法推导其递推公式;根据变系数估值序列的特点,运用灰色GM(1, 1)模型对变系数进行预测。Grey theory is widely used in the settlement prediction of road bed.The algebraic polynomials were used in this paper other than the common gray model to approach the discrete sequences.Grey algebraic curve model is put forward.The coefficients of a algebraic polynomial are considered to be variable,and the recursive formula of the coefficients in the model can be calculated by means of Laplace conditional extremum.On the basis of the characteristics of the varying coefficients' estimation series,GM(1,1)is used to predict the varying coefficients.

关 键 词：变系数 灰色代数曲线模型 GM(1 1) 沉降 预测 

分 类 号：TU416[建筑科学—岩土工程]