变异函数是克里金法中反映区域化变量空间变化特征的有效数学模型,但传统克里金方法中变异函数理论模型的选择和实验变异函数参数的设定具有一定的主观性.引入粒子群算法,对Kriging实验变异函数参数进行优化,提出了PSO-Kriging算法并结合实例进行三维建模.实验结果表明:PSO-Kriging算法与传统Kriging方法相比,误差降低29.14%,三维地质模型精度更高.
Variation function is an effective mathematical model reflecting the spatial variation of regionalized variables of Kriging.However,in traditional Kriging,model selection of variation function theory and parameter in the experimental variation function model are set with certain subjectivity.In this paper we introduced the particle swarm algorithm to optimize the parameters in experimental variation function of Kriging.We proposed PSO-Kriging algorithm with examples of three dimensional modeling.The experimental results show that the PSO-Kriging decreases error 29.14%and has higher precision of three-dimensional geological model.