电子学报 ›› 2019, Vol. 47 ›› Issue (3): 613-622.DOI: 10.3969/j.issn.0372-2112.2019.03.014

• 学术论文 • 上一篇    下一篇

基于自适应模拟退火的变频正弦混沌神经网络

胡志强1,2,3, 李文静1,3, 乔俊飞1,3   

  1. 1. 北京工业大学信息学部, 北京 100124;
    2. 泰山学院机械与建筑工程学院, 山东泰安 271000;
    3. 计算智能与智能系统北京市重点实验室, 北京 100124
  • 收稿日期:2017-08-28 修回日期:2018-01-12 出版日期:2019-03-25
    • 作者简介:
    • 胡志强 男,1988年7月出生,山东省泰安人,2013年毕业于桂林电子科技大学获硕士学位,现为北京工业大学信息学部博士研究生.主要研究方向为混沌动力学、混沌神经网络、污水处理建模与仿真、智能优化算法等.E-mail:zacharyhu33@163.com;李文静 女,1985年出生,陕西西安人,2013年7月毕业于中国科学院自动化研究所获工学博士学位,现为北京工业大学副教授,硕士生导师.主要研究方向为神经计算,人工神经网络,模式识别、神经科学.E-mail:wenjing.li@bjut.edu.cn;乔俊飞 男,1968年出生,内蒙古鄂尔多斯人.1998年于东北大学获博士学位,现为北京工业大学教授,博士生导师,主要研究方向为智能控制理论及应用,神经网络分析与设计等.E-mail:junfeiq@bjut.edu.cn
    • 基金资助:
    • 国家自然科学基金重点项目 (No.61533002); 国家自然科学基金青年科学基金 (No.61603009); 北京市科技专项课题领军人才 (No.Z1511000001315010); 北京工业大学日新人才计划 (No.2017-RX (1)-04)

Frequency Conversion Sinusoidal Chaotic Neural Network Based on Self-adaptive Simulated Annealing

HU Zhi-qiang1,2,3, LI Wen-jing1,3, QIAO Jun-fei1,3   

  1. 1. Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China;
    2. College of Mechanical and Architectural Engineering, Taishan University, Taian, Shandong 271000, China;
    3. Beijing Key Laboratory of Computational Intelligence and Intelligent System, Beijing 100124, China
  • Received:2017-08-28 Revised:2018-01-12 Online:2019-03-25 Published:2019-03-25
    • Supported by:
    • Key Program of National Natural Science Foundation of China (No.61533002); Youth Fund of National Natural Science Foundation of China (No.61603009); Science and Technology Project of Beijing Municipality -  (Leading Talents Program) (No.Z1511000001315010); New Talent Program of Beijing University of Technology (No.2017-RX (1)-04)

摘要: 针对变频正弦混沌神经网络寻优精度与收敛速度无法兼顾的问题,通过分析暂态混沌神经网络的优化机制和现有的退火策略,提出了一种基于自适应模拟退火策略的变频正弦混沌神经网络模型.该模型可以根据混沌神经元的Lyapunov指数来确定合适的自反馈连接权值.给出了混沌神经元的倒分岔图、Lyapunov指数及不同退火函数的时间演化图,证明了自适应模拟退火策略能够自主选择合适的退火速度,更有效的利用混沌全局搜索能力,并加快非混沌态的演化时间.为了证明该模型的有效性,将其应用于函数优化和组合优化问题中.仿真实验表明:(1)对于该模型退火速度的选择,自适应模拟退火策略比现有的几种退火方法更具有灵活性和适应性;(2)该模型在寻优精度和速度上比暂态混沌神经网络及其他改进模型具有更好的兼顾性.

关键词: 变频正弦混沌神经网络, Lyapunov指数, 自适应模拟退火, 优化计算

Abstract: The frequency conversion sinusoidal chaotic neural network (FCSCNN) cannot consider search accuracy and convergence speed simultaneously.In order to solve the mentioned problem,a novel self-adaptive simulated annealing (SSA) strategy is proposed by analyzing the optimization mechanism of the transiently chaotic neural network (TCNN) and the existing annealing strategy.It can give appropriate self-feedback connection weights based on the characteristics of Lyapunov exponent.The reversed bifurcation,Lyapunov exponent and annealing function evolution diagram of the chaotic neuron are given and the dynamic characteristic is analyzed.It shows that the SSA strategy can choose appropriate annealing speed in different stages,which can not only make full use of chaotic global searching ability but also accelerate convergence speed.Based on the neuron model,a novel FCSCNN with SSA strategy (FCSCNN-SSA) is proposed and applied to nonlinear function optimization and combinational optimization problems.The simulation results show that:(1) The SSA strategy can targeted choose the appropriate annealing speed,which is superior to other several existing simulated annealing methods for pertinence and adaptability and can be expanded to other similar models with same optimization mechanism; (2)FCSCNN-SSA can converge with a fast speed and search accuracy simultaneously than TCNN,TCNN-SEA,I-TCNN,NCNN,BFS-TCNN,FCSCNN.

Key words: frequency conversion sinusoidal chaotic neural network, Lyapunov exponent, self-adaptive simulated annealing, optimal computation

中图分类号: