模拟退火算法在概率积分法参数反演中的应用

Application of simulated annealing algorithm to parameters inversion in probability-integral method

针对传统方法求取概率积分法参数存在易发散且难以获得最优解的不足,本文提出将模拟退火算法应用到反演概率积分法参数中。利用模拟退火算法在搜索解时能够跳出局部最优解,获得全局最优解的特点,建立模拟退火法反演概率积分参数模型,求出概率积分参数。通过理论分析与实验验证,模拟退火算法反演概率积分法参数精度高,并对随机误差和粗差有一定的抗干扰能力,而且实验反演的地表下沉数据基本符合实测数据,获得了较好的结果,证明了该模型的可行性与有效性。

Simulated annealing algorithm is proposed to solve the divergence problem of parameter inversion in probability-integral method in this paper. Simulated annealing algorithm can jump out of the local optimal solution and achieve the global optimal solution,therefore,a simulated annealing algorithm model can be set up to get solution of the parameter inversion in probability-integral method. Through the theoretical analysis and experimental verification,simulated annealing algorithm is accurate in parameters inversion,and it is helpful to deal with both random error and gross error,and the result of experiment is similar to what has been measured. It proves that simulated annealing algorithm is feasible and effective in parameters inversion in probability-integral method.