东北大学学报:自然科学版 ›› 2015, Vol. 36 ›› Issue (11): 1558-1561.DOI: 10.12068/j.issn.1005-3026.2015.11.009

• 信息与控制 • 上一篇    下一篇

重力训练法

张鹏1, 李平2, 李青芮1   

  1. (1. 西北工业大学 自动化学院, 陕西 西安710000; 2. 辽宁石油化工大学 信息与控制工程学院, 辽宁 抚顺113000)
  • 收稿日期:2014-03-30 修回日期:2014-03-30 出版日期:2015-11-15 发布日期:2015-11-10
  • 通讯作者: 张鹏
  • 作者简介:张鹏(1983-),男,辽宁抚顺人,西北工业大学博士研究生; 李平(1964-),男,湖南涟源人,辽宁石油化工大学教授,博士生导师.
  • 基金资助:
    辽宁省自然科学基金资助项目(20102127); 辽宁省高校创新团队支持计划项目(2009T062.LT2010058).

Gravity Training Method

ZHANG Peng1, LI Ping2, LI Qing-rui1   

  1. 1.School of Automation, Northwestern Polytechnical University, Xi’an 710000, China; 2. School of Information and Control Engineering, Liaoning Shihua University, Fushun 113000, China.
  • Received:2014-03-30 Revised:2014-03-30 Online:2015-11-15 Published:2015-11-10
  • Contact: ZHANG Peng
  • About author:-
  • Supported by:
    -

摘要: 针对某些对象梯度不易计算的问题,提出一种新的优化训练方法.利用水往低处流的原理对参数进行寻优,将随机初始参数根据步长生成多维空间中点的坐标形式,用实际输出和目标输出之间的差来表示以此点为球心的球的半径,将最小半径的球作为下一步寻优的中心点.此方法具有无需计算梯度,初始值随机选取,易编程,寻优快等特点.通过仿真实验,此方法成功应用在PID控制器、LQR控制器和神经网络中,获取了最优参数,具有实际操作性.

关键词: 重力, 训练, 优化, 梯度, 神经网络, PID控制器, LQR控制器, 参数

Abstract: A new training method was proposed to solve the problem that some gradients of the objects are not easy to calculate. This method was based on the principle of gravity optimizes parameters. Random initial parameter based on step was set as coordinate form which in the midpoint of the multidimensional space. The error between the actual output and the target output was set as radius. This method had advantages which could not need to calculate the gradient and could randomly select initial value. This method was successfully used in the PID controller, LQR controller and neural network.

Key words: gravity, training, optimization, gradient, neural network, PID controller, LQR controller, parameter

中图分类号: