首页 | 本学科首页   官方微博 | 高级检索  
     检索      

改进量子粒子群BP神经网络参数优化及应用
引用本文:贾伟,赵雪芬.改进量子粒子群BP神经网络参数优化及应用[J].教育技术导刊,2019,18(10):30-35.
作者姓名:贾伟  赵雪芬
作者单位:宁夏大学新华学院 信息与计算机科学系,宁夏 银川 750021
基金项目:宁夏高等学校科学技术研究项目(NGY2017225)
摘    要:针对现有BP神经网络选取权值和阈值不精确问题,采用改进量子粒子群优化算法优化BP神经网络权值和阈值。首先在改进的量子粒子群优化算法中,采用双层多种群优化策略提高整个种群的寻优能力,然后在每个子群中使用混沌反向学习和Levy飞行增强子群寻优能力,最后利用改进的量子粒子群优化BP神经网络权值和阈值。实验结果表明,改进的量子粒子群优化算法能有效提高BP神经网络的全局寻优能力和收敛性,对数控高速铣削加工的铣削力进行准确预测。

关 键 词:量子粒子群优化  BP神经网络  多种群  混沌反向学习  Levy飞行  
收稿时间:2019-01-02

Parameters Optimization of BP Neural Network Using Improved Quantum-behaved Particle Swarm and Its Application
JIA Wei,ZHAO Xue-fen.Parameters Optimization of BP Neural Network Using Improved Quantum-behaved Particle Swarm and Its Application[J].Introduction of Educational Technology,2019,18(10):30-35.
Authors:JIA Wei  ZHAO Xue-fen
Institution:Faculty of Information and Computer Science, Xinhua College of Ningxia University, Yinchuan 750021, China
Abstract:Focused on the issue that the existing BP neural networks cannot effectively select weights and thresholds, an improved quantum-behaved particle swarm optimization is proposed to optimize the weights and thresholds of BP neural network. Firstly, in the improved quantum-behaved particle swarm optimization, a two-layer multi-swarm strategy is used to improve the global search ability. Secondly, chaotic opposition-based learning and levy flight are utilized in each sub-swarm to enhance the searching ability of sub-swarm. Finally, the improved quantum-behaved particle swarm optimization is used to optimize the weights and thresholds of BP neural network. Experimental results show that the proposed quantum-behaved particle swarm optimization can effectively improve the global optimization ability and convergence of BP neural network, and can accurately predict the milling force of NC high-speed milling.
Keywords:quantum-behaved particle swarm optimization  BP neural network  multi-swarm  chaotic opposition-based learning  Levy flight  
点击此处可从《教育技术导刊》浏览原始摘要信息
点击此处可从《教育技术导刊》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号