## Math/CS 143M, Numerical Analysis & Scientific Computing (Matrices), Fall 2017

Instructor: Plamen Koev

# New room for the rest of the semester: SH 346!

### New date for midterm 1: October 4, in class (Wednesday instead of Monday)

• Greensheet
• Lecture Notes
• Software used for demonstration in class.
• Student-run Facebook group from Fall 2016. Let me know if you form a new one for me to put the link here.
• Homework
1. Due September 18: CORRECTION: Solve all problems from chapter 1, sections 2, 3, and 4 of the lecture notes. Make sure to download the latest version of the notes.
2. NEW DATE: Due October 4: Solve all problems from chapter 1, sections 5, 7, and 8, and only problem 1 from chapter 1, section 6 of the lecture notes.
3. Due October 23. Solve all problems from Chapter 2, section 2, and Chapter 3, sections 1 through 6.
4. Due November 27. Solve all problems from Chapter 4, and Chapter 5, sections 1-7.
5. Due December 11 (Firm! There will be no extensions!): Solve problems in Chapter 5, section 8, problems 1, 2, 4, or problems 1, 4, 5.
• Practice exams for Fall 2017:
• Tasks you'll need to know how to perform in order to pass the class. These are in no particular order. Unless it says "on a computer" you should be able to do the tasks by hand.
1. Compute 1, 2, and infinity norms of vectors and of up to 3-by-3 matrices.
2. Calculate the cost of an algorithm both by hand an on a computer.
3. Generate matrices that subtract one row of a matrix from another, multiply a row by a number, or swap rows.
4. Generate (potentially complex) Givens rotations that create zeros in matrices.
5. Compute the LU, LDU, LU with partial pivoting, Cholesky, and QR decompositions on up to 3-by-3 matrices. Compute the LQ, QL, RQ decomposition on 2-by-2 matrices.
6. Perform 1 step of QR iteration with a shift on a 2-by-2 matrix.
7. On a computer: Write code that creates zeros in a matrix using elementary elimination operations or Givens rotations.
8. Understand properties of orthogonal matrices and be able to check if a matrix is orthogonal.