摘要
针对大坝安全监控中传统BP神经网络模型由于采用最速下降法求解网络权值而存在的计算过程复杂、易陷入局部极值点等缺点,提出大坝安全监控神经网络权值的协同粒子群优化求解方法。该方法先把网络权值的计算问题转化为粒子群的寻优问题,然后通过粒子群协同寻优实现对网络权值的计算。工程实例分析结果表明:基于协同粒子群算法的神经网络模型计算简单、收敛速度快、拟合精度高,为大坝安全监控分析提供了一种有效的新方法。
The calculation process of traditional BP neural network models employed in dam safety monitoring is complicated, and the calculations easily fall into local extreme points, due to use of the steepest descent algorithm in the calculation of BP neural network weights. In order to overcome these shortcomings and to optimize the weights of neural networks for dam safety monitoring, the cooperative particle swarm optimization algorithm is proposed. Then the calculation of neural networks weights is transformed into the cooperative optimization of particle swarms. The results of an engineering instance show that, with a simple calculation process, fast convergence and high precision, the neural network model based on the cooperative particle swarm optimization algorithm provides a new and effective method for dam safety monitoring.
出处
《水利水电科技进展》
CSCD
北大核心
2008年第4期8-10,75,共4页
Advances in Science and Technology of Water Resources
基金
国家科技支撑计划基金(2006BAC14B03)
国家自然科学基金(50579010)
国家重点基础研究发展计划(973计划)基金(2002CB412707)
关键词
协同粒子群
神经网络
大坝安全监控
cooperative particle swarm
neural network
dam safety monitoring