摘要
BP神经网络具有较强的学习能力,但在传统的研究中,隐含层节点、学习因子和动量因子往往采用试凑法得到相对较佳值,而试凑法在浪费较多时间的同时,可能得不到理想的BP神经网络输出,这对研究造成了一定的困难。文中采用智能算法来解决BP神经网络优化问题。遗传算法作为一种随机搜索算法,能够快速寻找到全局最优解,可以应用于本优化问题。因此,文章采用遗传算法优化BP神经网络上述参数,将改进后的BP神经网络运用于动力配煤非线性模型的研究。结果表明,采用遗传算法优化的BP神经网络具有较强的预测能力,对煤质的发热量预测误差优于线性平均模型误差,并且仿真表明动力配煤模型为近似线性的非线性模型,BP网络的输出值误差波动较小,结果理想。
The BP neural networls has strong learning ability,but in the traditional studies,hidden layer nodes, learning factors and momentum factors tend to use trial and error method to get relatively better value,at the same time,trial and error method in a waste of more time,the BP neural network output may not be ideal,this caused some difficulties to research. In this paper,the intelligent algorithm is applied to solve the problem of the BPneural network optimization. Genetic algorithm as a kind of random search algorithm,it can find the global optimal solution quickly, can be applied to the optimization problem. This paper uses genetic algoritlim to optimize the BP neural networls parameters,and ap-plies the improved BP neural networls in the study of nonlinear model of power coal blending. The results showthat using genetic algorithm to optimize the BP neural networls has strong ability of prediction,calorific value of coal quality prediction eror is supemodel,and the simulation results show the approximate linear dynamic coal blending model of nonlinear modwork error is less volatile,the result is ideal.
作者
李吉朝
张海英
王惠琴
Li Jichao Zhang Haiying Wang Huiqin(Key Laboratory of Complex System Control and Intelligent Information Processing,Xi ＇an University of Science and Technology,Xi ＇an 710048,China Xi＇ an Environmental Monitoring Station, Xi ＇an 710048,China)
出处
《微型机与应用》
2017年第9期60-63,66,共5页
Microcomputer & Its Applications
关键词
动力配煤
BP神经网络
遗传算法
非线性
power coal blending
BPneural network
genetic algorithm
nonlinear
