Abstract: Matching the parameter of theoretical model of variogram is one of the specific basic content in geostatistics. Aiming at the disadvantages of conventional used methods in match theoretical variogram, in order to match theoretical variogram automatically, based on the quick convergence of Chaos Particle Swarm Optimization（CPSO） algorithm used in solving nonlinear programming problems, and ergodicity, global optimization of chaos, a new match method of theoretical variogram is performed. Spherical model, which is transformed into minimum problem, is used in the experiment. CPSO is used to find the satisfactory fitness value. Simulations show that the new algorithm has powerful optimizing ability, higher precision, and can automatically match the theoretical variogram effectively.

Key words: geostatistics, Particle Swarm Optimization（PSO） algorithm, theoretical variogram, spherical model, fitting

摘要： 理论变差函数的参数拟合是地质统计学特有的基本内容之一。针对现有的拟合方法的不足之处，为了快速实现理论变差函数参数的自动拟合，充分利用混沌粒子群优化算法在求解非线性优化问题时的快速收敛性，混沌运动的遍历性和全局寻优的特点提出了实验变差函数的混沌粒子群自动拟合算法。选用一阶球状模型，并将其转化为极小值问题，利用混沌粒子群优化算法找到满意的适应值。实验结果表明，该方法寻优能力强，精度高，能有效实现参数的自动拟合。

关键词: 地质统计学, 粒子群优化算法, 变差函数, 球状模型, 拟合