AMATH 352
Autumn Quarter, 2003

Applied Linear Algebra and Numerical Analysis

Important Announcements and Hints

12/10/03: Office hours Monday 12/15 and Tuesday 12/16: 11:30-1:30 in MSCC.
12/4/03: Typo on Final project, Number 1(a): the exact solution was typed incorrectly, it is y(t) = t - 1/(t-1).
12/1/03: Typo on Homework 8, Number 1, just below the integral: The points are x_1 = -xbar and x_2 = xbar.
12/1/03: Typo on Homework 8, Number 1(a): in the Hint, the linear function should be a_0 + a_1 x, not a_1 + a_1 x.
12/1/03: Today's handout on adaptive quadrature and the m-file.
12/1/03: LeVeque will be out of town at a meeting this Thursday and Friday, Dec 4-5. Pelanti will lecture on Friday and will hold extended office hours at MSCC, Thursday 3-5pm and Friday 9-10am. LeVeque will have office hours Wednesday 1:30-2:30 in GUG 408A (as usual) and Wednesday 3-4 at MSCC.
11/24/03: LeVeque will have office hours at the MSCC lab Tuesday 11/25, 10-11am. Pelanti will be there Tuesday 1:00-2:30, and will not have office hours on Wednesday this week.
11/19/03: There will be a review session Thursday 11/20 at 4:30pm in Guggenheim 410. No calculators on the second midterm.
Review sheet
11/11/03: Reminder: The second midterm is Friday 11/21 and will cover through this week's material.
11/11/03: The handout from yesterday is available on the webpage and here:
- Some notes on least squares (11/10): .ps ... .pdf
10/29/03: Reading for Friday: pp. 104-112, 122-126.
10/29/03: We will not have office hours in the MSCC lab this week since there is no homework due. I will have office hours W 1:30-2:30 and Th 9-10 in GUG 408A.
10/28/03: My office hours this morning 9-11 are in Guggenheim 408A.
10/27/03: The midterm review session will be Monday 10/27, 4:30 - 6:00pm, in Johnson 119.
10/22/03: Office hours next week: Monday 10/27, 3:30-4:30 and Tuesday 10/28, 9-11.
10/22/03: HW4 hint:
For #1 (Section 2.2 #19), plot the iteration function g(x) to get an idea of how this iteration behaves and what's being asked. For part (a) you can use any interval [sqrt(2), b] that contains x_0. An easy way to establish the inequalities needed for parts (a) and (b) is to note that 0 < g'(x) < 1/2 for x > sqrt(2) and g'(x) < 0 for 0 < x < sqrt(2) and use the mean value theorem.
10/21/03: HW4 corrections:
#2 f(x) = x^2 - 2, not f(x) = x^2 - 1.
#5 x = A\b makes more sense (though b = A\b would work fine but overwrites the right-hand side b with the solution of the system)
10/21/03: A sample code fpi1.m for the fixed point iteration problem of HW3 is now in the hw3 directory. This may be useful in writing a program for Newton's method if you have been having problems with this.
This Thursday 10/23 I will not be able to make my usual office hours 9-10. Instead I will be in the MSCC lab 11-12 on Thursday. Sorry for any inconvenience.
Midterm 1 will be next Wednesday, October 29. As announced on the information sheet handed out the first day (and on the webpage), you may bring one page of handwritten notes (8.5x11 page, both sides) to the exam. You may bring a calculator. I will hand out a review sheet.
I also plan to have a review session before the exam, tentatively scheduled for Monday 10/27, 4:30 - 6:00pm.
10/16/03: As I announced yesterday in AMath 352, I'm planning to switch to blackboard lectures for a while anyway, so printed slides will not be available. Please come to class and take notes!
10/16/03: There was a questions about HW3 #6, and whether the program should give exactly the same number of iterations as the estimate. No, the estimate should be close but is only an estimate. Why might it not give the exact number?
10/13/03: If you're interested in attending graduate school, there's a panel discussion on October 21 that may interest you.
10/13/03: Read Sections 6.1 - 6.3 for this week's lectures. Additional material on linear algebra will also be discussed.
10/13/03: Marica's office hours this week will be Tuesday 2:30 - 4:00 and Wednesday 3:30 - 5:00, with none on Thursday.
10/8/03: Homework #2:
A correction to problem 2(c): the last term in the expressions of x_1, x_2 should be epsilon^2 (instead of epsilon^3).
Clarification of Hint on Section 1.1 #9(a): "observe that this is an increasing function of x..." I should have said "observe that the expression for the error bound is an increasing function of xi ..." (where xi is the greek letter in this problem).
Note on Section 1.1 #14: Note that P(x) = x is the first-order Taylor polynomial approximating sin(x). But it is also the second-order Taylor polynomial approximating sin(x). (Why?) So it's possible to use either of two different remainder terms to approximate the error. Which gives a better estimate of the actual error?
10/6/03: I'm not on jury duty this week.
10/3/03: Type on Slide p. 10 from Lecture 1F: 400 should be 200 in the Taylor series for exp(-20).
10/1/03: Please read sections 1.1 and 1.2 in the book and review the basic calculus and Taylor series material. I'll be talking about Sec. 1.2 on Friday. You might also want to read Sec 1.3, as some material from this section will be used later (with some discussion in lecture). Next week we will start Chapter 2.
9/30/03: Tim Reluga, an AMath grad student with extensive matlab and 352 experience, will be working as an MSCC Lab Assistant this quarter and will be available certain hours as another resource. This week his hours are Thursday 1:30 - 3:30. Check the course webpage or the MSCC webpage for hours from next week.
9/30/03: There will be two introductory sessions in the MSCC lab this week: Wednesday at 4:30pm and Thursday at 12:30pm, in CMU B027. I will demonstrate how to use some basic features.