摘要
针对BP人工神经网络具有易陷入局部极小等缺陷,提出了将遗传算法与神经网络结合,同时优化网络结构的权值与阈值的思想,建立了基于遗传算法的围岩松动圈预测的神经网络模型。用该模型对巷道围岩松动圈厚度进行了预测并与BP预测结果相比较。结果表明,该遗传神经网络模型可靠,预测精度高,用来对围岩松动圈厚度进行预测是有效的和可行的。
Considering some defects of BP Neural Network, the idea that the power size and the threshold value of the network structure is optimized by combining genetic algorithm with neural network is presented. Based on genetic algorithm, the prediction model of loosen zone around roadway is built. Finally, the prediction on the thickness of the loosen zone around roadway is made with this GA- BP model, and the results are compared with the BP predicting results. The result shows that the GA- BP model is reliable and precise, and it is effective and feasible to predict the thickness of the loosen zone around roadway.
出处
《岩土工程技术》
2006年第5期237-239,266,共4页
Geotechnical Engineering Technique
关键词
遗传算法
松动圈厚度
预测
围岩
genetic algorithm
thickness of the loosen zone
prediction
surrounding rock mass
