Research Advice for Graduate Students

Research Advice for Graduate Students

Professor Ryan Loxton, BSc (Hons), PhD, FAustMS, FORS || School of Electrical Engineering, Computing, and Mathematical Sciences || Curtin University, Australia

Time and classroom is to be announced.

Abstract: Graduate students face the daunting task of transitioning from “learner” to “inventor” and this requires much more than just technical knowledge – for example, the ability to write and communicate is crucial, but such “soft skills” are often neglected in the education system. This talk aims to address this gap by providing graduate students with advice on improving academic writing and building a long-term research career. The speaker will discuss such questions as: “How to choose a research direction?”, “How to do high-quality research?” and “How to write publications?”. The talk is based on the speaker’s personal experience as a senior researcher who has worked with many graduate students in China.

Ryan Loxton is a professor in the School of Electrical Engineering, Computing, and Mathematical Sciences at Curtin University. His work focuses on using advanced mathematics to optimise complex processes in a wide range of applications such as mining, oil and gas, agriculture, and industrial process control. Ryan is a passionate advocate for industry engagement and has worked with many companies including Woodside Energy, Linkforce, and Vekta Automation. He has received several prestigious awards for his work, including the 2019 JH Michell Medal from the Australian Mathematical Society and the 2014 West Australian Young Scientist of the Year Award. Ryan leads the optimisation theme in the new Australian Centre for Transforming Maintenance through Data Science, which is co-funded by the Australian Research Council and industry partners Alcoa, BHP Billiton, and Roy Hill. He has published over 60 journal articles, including papers in Automatica, IEEE Transactions on Automatic Control, and SIAM Journal on Control and Optimization.

 


Short Course – Dynamic Optimization and Computational Optimal Control

Title: Dynamic Optimization and Computational Optimal Control

Professor Ryan Loxton, BSc (Hons), PhD, FAustMS, FORS || School of Electrical Engineering, Computing, and Mathematical Sciences || Curtin University, Australia

Time and classroom information: Institute of CSC Room 304, 18:30 – 21:00 // 智能系统与控制研究所304教室

This short course will introduce students to the computational tools needed to solve dynamic optimization and optimal control problems. Two core themes will be discussed: variational methods for gradient computation and switching time optimization techniques for handling switching dynamics. These concepts can be combined to solve a wide variety of constrained optimal control problems with non-standard features such as time-delays, mode switches, and state impulses.

The course schedule is as follows. (浙大玉泉校区智能系统与控制研究所304教室,时间为18:30 – 21:00)

  • 19 April, 2 hours (night): Optimal parameter selection problems
  • 22 April, 2 hours (night): Switching time optimization
  • 23 April, 2 hours (night): Optimal control problems (时间可能需要调整,稍后通知)

Ryan Loxton is a professor in the School of Electrical Engineering, Computing, and Mathematical Sciences at Curtin University. His work focuses on using advanced mathematics to optimise complex processes in a wide range of applications such as mining, oil and gas, agriculture, and industrial process control. Ryan is a passionate advocate for industry engagement and has worked with many companies including Woodside Energy, Linkforce, and Vekta Automation. He has received several prestigious awards for his work, including the 2019 JH Michell Medal from the Australian Mathematical Society and the 2014 West Australian Young Scientist of the Year Award. Ryan leads the optimisation theme in the new Australian Centre for Transforming Maintenance through Data Science, which is co-funded by the Australian Research Council and industry partners Alcoa, BHP Billiton, and Roy Hill. He has published over 60 journal articles, including papers in Automatica, IEEE Transactions on Automatic Control, and SIAM Journal on Control and Optimization.


最优控制导论教学日历-更新2018年4月25日

第一讲:课程导论(2018年3月5日,第一周)

  • 什么是最优控制问题
  • 最优控制有哪些应用
  • 最优控制的数学本质
  • 课程参考书籍与资料

第二讲:古典变分方法简介(2018年3月7日,第一周)

  • 泛函极值问题的例子
  • 欧拉折线近似解法
  • 拉格朗日变分解法
  • 欧拉-拉格朗日方程

第三讲:古典变分法应用-经典力学(2018年3月12日,第二周)

  • 广义坐标
  • 拉格朗日力学
  • 哈密顿力学

第四讲:带微分/积分方程约束的泛函极值问题-I(2018年3月14日,第二周)

  • 回顾无约束最优化问题
  • 回顾带等式约束最优化问题
  • 带微分方程约束的泛函极值问题的变分方法

第五讲:带微分/积分方程约束的泛函极值问题(2018年3月19日,第三周)

  • 最简单的最优控制问题
  • 带积分方程约束的泛函极值问题
  • 三类最优控制指标的分类与等价性

第六讲:横截条件:I(2018年3月21日)

  • 终端时刻固定,终端状态自由
  • 终端时刻自由,终端状态固定

第七讲:横截条件:II(2018年3月26日、28日-请假调课,2018年4月2日)

  • 终端时刻和状态自由且无关
  • 一般情况的横截条件
  • 光滑最优控制问题

第八讲:线性二次最优控制 I(2018年3月26日、28日-请假调课,2018年4月4日)

  • 无约束最优控制例子
  • 线性二次最优控制

第九讲:线性二次最优控制 II(2018年4月9日)

  • 最小能量控制问题(Robert L. Williams II and Douglas A. Lawrence, Linear State-Space Control Systems)
  • 线性系统的跟踪问题
  • 带终端约束的反馈控制

第十讲:内点条件(2018年4月13日,晚上8点45分~10点,CSC304,调课补课

  • 内点约束问题
  • 内点条件推导
  • 小车折返问题

第十一讲:最大值原理及应用 (2018年4月16日)

  • 最大值原理的描述
  • 时间最优控制问题

第十二讲:最优性原理与动态规划(2018年4月18日)

  • 多级决策问题的最优性原理概述
  • Bellman方程
  • 运用动态规划求解离散系统最优控制

第十三讲:离散化与遍历方法(2018年4月20日,晚上8点45分~10点,CSC304,调课补课

  • 遍历离散状态空间
  • 遍历离散状态空间与控制空间
  • 近似值函数
  • 维数灾难
  • 离散时间系统的策略迭代与值迭代方法

第十四讲:动态规划求解连续最优控制(2018年4月23日)

  • Hamilton-Jacobi-Bellman方程推导
  • 求解连续最优控制问题示例
  • 连续时间自适应动态规划(策略迭代方法)

第十五讲:动态规划求解线性二次(LQ)最优控制(2018年4月25日)

  • 离散系统LQ问题
  • 连续系统LQ问题
  • 滚动时域优化控制

第十六讲:最优控制的计算方法(2018年4月27日,调课,兴趣讨论,随意参加

参考资料

  • Desineni Subbaram Naidu, Optimal Control Systems, CRC Press
  • 张杰、王飞跃,最优控制,清华大学出版社
  • Mark Kot, A First Course in the Calculus of Variations, AMS
  • Robert L. Williams II and Douglas A. Lawrence, Linear State-Space Control Systems, John Wiley & Sons, INC.

Dynamics – Fall 2017

浙江大学现代教务网记录:控制工程科学前沿 秋{第1-8周|3节/周} 许超/周建光 玉泉教7-304(多) [双语]

本门课程针对自动化专业本科生,主要讲授经典动力学的基本内容,为机器人、无人系统等方面深入研究提供必要的物理基础。

Lecture 1 – Sept. 19, 2017

  • 课程简介,包括:为什么学习动力学;本课程涵盖哪些方面的内容;课程考核方式(平时作业、课程项目、考勤、期末大作业);课程微信群;实验室有关的实验设施;
  • 向量代数复习;
  • 作业布置(9月26日交);
  • 讲义PDF文件通过微信发布;

Lecture 2 – Sept. 26, 2017

  • 小结向量代数复习内容;
  • 二维平面坐标系变换;
  • 三维空间坐标系变换;
  • 补充内容:向量和伪向量、坐标变换的分量推导方法(往年课件);
  • 作业布置(10月10日交);

Lecture 3 – Oct. 10, 2017

  • 随堂测试;
  • 复习坐标变换;
  • 证明欧拉转动定理(代数证明方法);
  • Kane简单转动的证明和向量表示公式;
  • 转动轴方向和转动角度的计算;
  • 作业布置(10月17日交)

Lecture 4 – Oct. 17, 2017

  • 复习题目:万向节;
  • 无穷小转动与微分方程;
  • 刚体运动学(动系下的速度、加速度计算公式);

Lecture 5 –

Lecture 6 –

Lecture 7 – Nov. 12, 2017

  • 时间改动到周日,教十八304教室
  • 本学期最后一次课,由于国庆节放假少了一次课
  • 变分法及分析力学初步(没有讲到广义力部分)
  • 需要给学生补充材料,参数共振、Classical Dynamics的教材
  • 作业布置