摘要
基于目前最优化方法理论教学与实验教学脱节的现状,设计了一套实验课程,包括基本算法和课程项目两大模块,分别覆盖了经典的最速下降法、Newton法、拟Newton法、共轭梯度法、惩罚函数法和近年来广泛使用的随机梯度下降法、支持向量机算法等。该实验课程能有效衔接最优化方法的理论与实验教学,有助于学生深入理解最优化理论及思想、掌握最优化算法设计技巧、提升创新能力和工程实践能力。
Based on the development of optimization theory and algorithm,this paper proposes a program design about the experimental courses.Classical optimization algorithms including steepest descend method,Newton method,quasi-Newton method,conjugate gradient method,penalty function method and the state-of-the-art algorithms such as stochastic gradient descend algorithm and support vector machine method are covered by the experimental courses.The experimental courses designed in this paper will help the learners deeply understand the optimization theory,master the design trick of optimization method,cultivate and improve their innovation capacity and engineering practice capacity.
作者
渐令
孙清滢
邵红梅
梁锡军
高卫峰
JIAN Ling;SUN Qingying;SHAO Hongmei;LIANG Xijun;GAO Weifeng(College of Science,China University of Petroleum,Qingdao 266580,Shandong,China)
出处
《实验室研究与探索》
CAS
北大核心
2018年第10期223-225,309,共4页
Research and Exploration In Laboratory
基金
中国石油大学教学改革项目(JY-B201857
QN201822
JY-A201403)
关键词
最优化方法
实验课程
优化算法
项目设计
optimization method
experimental curriculum
optimization algorithm
project design
