【摘要】 针对厚松散层薄基岩矿区开采沉陷变形预计中存在的下沉系数偏大问题,以23个开采工作面的地质采矿条件及其下沉系数为学习和测试样本,将文化-随机粒子群算法(CA-rPSO)和支持向量机(SVM)相结合,利用CA-rPSO的快速并行寻优功能优化SVM参数,建立了厚松散层薄基岩开采条件下下沉系数的SVM预测模型。通过实例验证SVM的预计结果与实际符合较好。更多还原

【Abstract】 In order to resolve the problem that subsidence coefficient is too large in predicting subsidence and deformation when mining under thick loose bed and thin bed rock,taking the geological and mining conditions and subsidence coefficients from 23 mining faces,combining culture-random particle swarm optimization with SVM so as to optimize SVM parameters with fast parallel optimizing function,SVM subsidence coefficient prediction model for mining under thick loose bed and thin bed rock was set up.Examples showed that prediction value well met actual result.更多还原