结合Markov模型和COSGSIM的插值方法

An Interpolation Method Using the Markov Model and COSGSIM

针对协同序贯高斯模拟在确定插值公式中的权值时出现的交叉矩阵不稳定问题,提出一种基于Markov模型和协同序贯高斯模拟的插值方法.利用Markov模型的屏蔽效应,假设待模拟硬数据可以屏蔽其位置以外的硬数据对待模拟位置处软数据的影响,对协同序贯高斯模拟进行逼近时只保留与待模拟硬数据同位置的软数据,而不考虑其他位置的软数据,从而解决了交叉矩阵不稳定的问题.实验结果证明,该方法插值误差小,模拟效果好.

An interpolation method exploiting both the Markov model and sequential Gaussian co-simulation（COSGSIM） is proposed to solve the instability of cross matrix when determining the weights of the interpolation equation in COSGSIM.According to the hypothesis of screening effect provided by a Markov model,the hard datum to be simulated screens the influence of any other hard data on the soft datum at the same location.Therefore,the instability of cross matrix can be dissolved by preserving only the soft datum that has the same location with the hard datum to be simulated,while ignoring the impact by other soft data when approximating the COSGSIM.Experimental results show that the deviation is small and the simulated result is good.