遗传算法求概率积分法预计参数程序设计

Program Design for Predicting Parameters Calculation of Probability Integral Method Based on Genetic Algorithm

目前基于概率积分法求取地表移动预计参数的优化算法,如线性最小二乘法,具有易发散、易出现局部解、对初值要求较高、抗粗差干扰性弱等问题。遗传算法(Genetic Algorithm)是一种高效的全局寻优搜索方法,采用人工进化的方式对目标空间进行随机优化搜索。采用遗传算法,利用地表任意点的概率积分法数学模型编写了求取地表移动预计参数的程序,通过皖北某矿井的1013工作面的观测数据并结合地质采矿条件反演了地表移动预计参数,通过与传统的优化算法所求取的参数比较,证明了遗传算法反演结果的准确性和可靠性。

The optimization algorithm based on the probability integral method was used to calculate the parame- ters of surface movement deformation, such as linear least squares, which has several problems of easily diver- gent results, being prone to local solutions, strong dependence on the initial values, and weak ability of anti - interference. Genetic Algorithm is a kind of efficient global algorithm for searching optimal solution by using arti- ficial evolution system. Program based on this kind of algorithm was designed, by which the parameters of sur- face movement deformation can be calculated by using the arbitrary point observation data and geological mining conditions of the working face. The results of experiments in practice proved feasibility and practicability of the method.