概率积分法预计参数的智能优化选择方法研究-中国管理科学.pdfVIP

第30 卷 第3 期 采矿与安全工程学报 Vol.30 No.3 2013 年05 月 Journal of Mining Safety Engineering May 2013 文章编号：1673-3363-(2013)03-0385-051 概率积分法预计参数的智能优化选择方法研究 卞和方，杨化超，张书毕 ( 中国矿业大学国土资源与灾害监测国家测绘地理信息局重点实验室，环境与测绘学院，江苏 徐州 221116) 摘要 为有效确定概率积分法预计参数，提高预计值的精度。将粒子群优化(PSO)算法和BP 神经 网络进行融合，采用改进的混合粒子群优化算法优化神经网络的权值和阈值。在分析概率积分法 参数与地质采矿条件之间关系的基础上，建立了基于PSO 优化BP 神经网络的概率积分法预计参 数的优化选择模型。以我国典型的地表移动观测站资料作为网络的学习训练样本和测试样本，将 计算结果与实际值进行了对比分析，并与改进BP 算法的计算结果进行了比较。结果表明，PSO-BP 神经网络方法用于概率积分法预计参数的选取是可行的，收敛速度更快，计算精度更高。 关键词 地表移动；概率积分法；粒子群优化算法；BP 神经网络；优化选择 中图分类号 TD 163.3 文献标志码 A Research on intelligent optimization for predicting parameters of probability-integral method BIAN He-fang ，YANG Hua-chao ，ZHANG Shu-bi (NASG Key Laboratory of Land Environment and Disaster Monitoring ，School of Environment Science and Spatial Informatics，China University of Mining Technology ，Xuzhou ，Jiangsu 221116 ，China) Abstract In order to effectively determine the prediction parameters of probability-integral method and to improve the prediction accuracy, a new method by combining Particle Swarm Optimization (PSO) algorithm and BP neural network (PSO-BP) was presented. In this method, an improved hybrid PSO algorithm was used to optimize the connection weights and thresholds values of BP neural network. An optimization model for prediction parameters of probability-integral method using this hybrid PSO-BP neural network algorithm was constructed based on analyzing the relationship between the parameters and geological mining conditions. Typical data of surface moving observation stations was used as learning and test samples. Analysis was made