數值最佳化與應用

This course is about how to numerically solve optimization problems. When you try to minimize the energy in a physical system, train a machine learning model, or optimize a design, numerical optimization plays an important role. In this course, we will start with the basic methods such as gradient descent and Newton’s method—and work our way up to powerful techniques like quasi-Newton methods and interior-point algorithms. The focus is on understanding how and why these methods work, and also when to use which method. You will get hands-on experience implementing algorithms, analyzing their performance, and applying them to real-world problems.

Calculus、Linear Algebra and Computational mathematics

we will use blackboards

homework assignments will be given biweekly and two written exams will be given. The grading policy is 60 % for homework assignments, and 25 % for the higher score in exams, 15 % for the lower one.

Numerical Optimization, 2nd Edition, Jorge Nocedal & Stephen J. Wright

地點

SA R216

時間

make appointments by emails

聯絡方式

mingcheng.shiue@gmail.conm