| BP网络的改进及模拟退火神经网络在地学中应用 |
| |
| 引用本文： | 席道瑛,张涛.BP网络的改进及模拟退火神经网络在地学中应用[J].物探化探计算技术,1996,18(3):219-224. |
| |
| 作者姓名： | 席道瑛 张涛 |
| |
| 作者单位： | 中国科学技术大学地球和空间科学系 |
| |
| 摘 要： | 本文针对ＢＰ神经网络收敛速度慢的缺点，提出了改进方案。还引进了模拟退火理论，将模拟退火算法与ＢＰ神经网络计算方法结合，并通过对正弦函数岩性识别及跨孔地震层析成象的试验，证明模拟退火ＢＰ算法是有效的、大范围收敛的。而且，具有良好的收敛速度和很高的精确度。
|
| 关 键 词： | BP神经网络 收敛速度 模拟退火 |
AN IMPROVEMENT ON BP NEURAL NETWORK AND THE APPLICATION OF THE THEORY OF SIMULATED ANNEALING TO GEOSCIENCES |
| |
| Xi Daoying,Zhang Tao.AN IMPROVEMENT ON BP NEURAL NETWORK AND THE APPLICATION OF THE THEORY OF SIMULATED ANNEALING TO GEOSCIENCES[J].Computing Techniques For Geophysical and Geochemical Exploration,1996,18(3):219-224. |
| |
| Authors: | Xi Daoying Zhang Tao |
| |
| Institution: | Univ. of Sci. and Tech of China |
| |
| Abstract: | This paper puts forward an improvement plan to counter the slowness of convergence rate of BP artificial neural network. It also introduces the theory of simulated annealing neural network. Through the experiments of lithologic characters distinguishment of sine function and the tomography of crosswell seismic, it proves that the simulated annealing BP algorithm is valid, and has a large range of convergence, good convergence rate and high accuray. |
| |
| Keywords: | BP neural network convergence rate simulated annealing |
| 本文献已被 CNKI 维普 等数据库收录！ |
