双变量散点图的曲线拟合方法及其应用

提出了一种基于退火技术来获取优化的拟合模型曲线的方法，基于该方法编制出双变量散点图曲线拟合程序的FITBIVAR。该程序的输入参数通过参数文件来设置，并遵照GeoEAS格式，根据实际应用特点，它提供变量的对数变换和结果反变换，FITBIVAR程序建立孔隙度，声波时差数据对和孔隙度，渗透率数据对的关系模型。结果表明，基于线性模型预测的孔隙度与岩心孔隙度之间的绝对误差为0．904％，相对误差为5．8％，基于FITBIVAR模型预测的孔隙度与岩心孔隙度之间的绝对误差为0．842％，相对误差为5．28％，基于FITBIVAR模型解释的渗透率与岩心参透率的相关性明显好于线性模型解释的相关性。

Curve Fit for Bivariate Scatter Plot with Simulated Annealing Technique and Its Application

An optimal fitting method of modeling curves is presented based on a simulation annealing technique and a fitting program FITBIVAR of bivariate scatter plot is developed by using this method In this program the input parameters are set by parameter files written in GeoEAS format and based on the actual application,the logarithm transform of variables and inverse transform of the results are provided According to the program FITBIVAR,relation models between porosity and sonic moveout as well as porosity and permeability are built up The results show that the absolute error between predicted porosity and core porosity is 0 904%, relative error is 5 8% when a linear model is used When FITBIVAR model is used the absolute error is 0 842%,relative error is 5 28% This demonstrates that the correlativity of predicted porosity and core porosity based on the model FITBIVAR is better than the linear model