基于蚁群算法的开采沉陷计算参数反演

Calculation parameters inversion of mining subsidence based on ant colony algorithm

针对开采沉陷计算参数反演方法存在早熟收敛现象、容易陷入局部最优解的问题,提出将蚁群算法运用于开采沉陷计算参数反演:首先对反演参数的搜索空间进行离散,将参数反演问题转化为组合优化问题,然后建立了基于最大-最小蚁群算法的概率积分法计算参数反演流程。应用实例表明,蚁群算法对观测站测点缺失具有较强的抗干扰能力,较最小二乘法和模矢法有较好的拟合效果。

For problems of premature convergence and easily fall into local optimal solution existed in inversion method of calculation parameters of mining subsidence, ant colony algorithm was applied to calculation parameters inversion of mining subsidence. Firstly, search space of inversion parameter is discretized, thus parameter inversion problem is transformed into combination optimization problem, and then flow of calculation parameters inversion of probability integral method based on max-min ant colony algorithm is established. Application results show that the ant colony algorithm has strong anti- interference ability for missing data of measuring points, which has better fitting effect than least square method and pattern search method.