【作者】 金菊良； 魏一鸣；

【Author】 JIN Ju-liang ,WEI Yi-ming (I.Hefei University of Technology,Hefei 230009,China2.Institute ofPolicy&Management,Chinese Academy of Sciences,Beijing 100080,China)

【机构】 合肥工业大学； 中国科学院科技政策与管理科学研究所；

【摘要】 考虑到地统计学中实际变异函数曲线上各点的重要性和计算精度各异,文中给出了在变异函数实际值与理论模型计算值之残差的加权平方和最小准则下估计变异函数参数的新方法,实质上这是一个非线性优化问题,可用我们提出的加速遗传算法(AGA)进行求解。实例的计算结果表明,AGA较常规方法简便、通用,而且估计精度高,它可直接在目标函数值最小准则下同时估计块金值、变程、基台值等变异函数的参数,在处理各种复杂的空间优化问题中具有广泛的应用价值。更多还原

【Abstract】 :The importance and computational precision of different points is not same on the practical variogram curve in geostatistics.For this reason,in this pape~a method is suggest~to estimate the variogram parameters such as range,nugget and sill at the same time by using the accelerating genetic algorithm(AGA) developed by the authors,under the weighted least square criterion of e~ors between practical vailogram values and theoretical model computational values.The results of the application example show that AGA is convenient,general and high precision.AGA can be widely applied to different spatial optimal problems in geostatistics更多还原