基于改进Elman神经网络的徽派古建筑寿命预测

发布时间：2018-04-14 00:32

  本文选题：Elman神经网络 + 粒子群　； 参考：《中国科学技术大学学报》2017年10期

【摘要】：徽派建筑是我国四大古建筑流派之一,木构件是徽派建筑的核心.准确预测徽派木构件的寿命,对于古建筑的保护具有重要的意义.目前系统考虑多种因素对木构件寿命共同影响的研究较少,Elman神经网络是一种典型的多层动态递归神经网络,通过存储内部状态使其具备映射动态特性的功能,从而使系统具有适应时变特性的能力,可用于预测木构件复杂的非线性时变系统的建模.针对基本的Elman神经网络存在训练速度慢、容易陷入局部极小值的特点,使用带有自适应变异算子的粒子群优化算法对基本的Elman神经网络进行改进,优化网络中各层之间的连接权值,提高学习速度,并在全局范围内寻找最优解.仿真结果表明,改进后的网络能较准确地拟合训练值,并进行有效预测,能够较好应用于徽派古建筑寿命预测.
[Abstract]:Huizhou architecture is one of the four ancient architectural schools in China, and wooden components are the core of Huizhou architecture.It is of great significance for the protection of ancient buildings to accurately predict the life of Huizhou wooden components.At present, there are few researches on the influence of many factors on the life of wood components. Elman neural network is a typical multi-layer dynamic recurrent neural network, which can map dynamic characteristics by storing internal states.Thus, the system has the ability to adapt to the time-varying characteristics and can be used to predict the modeling of complex nonlinear time-varying systems with wood components.The basic Elman neural network has the characteristics of slow training speed and easy to fall into local minimum. The particle swarm optimization algorithm with adaptive mutation operator is used to improve the basic Elman neural network.The connection weights of each layer in the network are optimized, the learning speed is improved, and the optimal solution is found in the global scope.The simulation results show that the improved network can fit the training value accurately and predict effectively, and it can be applied to the prediction of the life of Huizhou ancient buildings.
【作者单位】： 安徽建筑大学电子与信息工程学院;安徽建筑大学机械与电气工程学院;
【基金】：十二五国家科技支撑计划(2012BAJ08B00) 安徽质量工程项目(2014zdjy091) 安徽建筑大学博士启动基金 易海人才工程资助
【分类号】：K879.1;TP183
，

本文编号：1746936