MATH662: Quiz 1
Answer the following questions in no more than 10 minutes.
1. Do n linearly independent vectors v1,
., vn in a vector space (V , R, +) form a basis?
2. Is the set of 2 2 diagonal matri
MATH662 Midterm Examination 03/06/14
Answer the following questions explaining all steps that lead to a solution.
1. Let A Cmm. Use the singular value decomposition of A to nd the eigenvalue decomposi
<TeXmacs|1.0.7.18>
<style|<tuple|generic|maxima>
<\body>
<section|MATH662 Lecture 4 (1/21/14)>
<subsection|Projection>
<\remark>
One way to mathematically express the very valuable concept of dimensio
<TeXmacs|1.0.7.19>
<style|generic>
<\body>
<section|MATH662 Homework 6, Issued 3/28/14, Due 4/3/14>
1. In eigenproblem algorithms we often reduce the size of a matrix after
finding one of its eigenval
<TeXmacs|1.0.7.19>
<style|generic>
<\body>
<section|MATH662 Homework 7, Issued 4/3/14, Due 4/10/14>
1. Problem 36.1 from Trefethen & Bau.
<em|Solution>. At step <math|n> the incomplete factorization o
<TeXmacs|1.0.7.18>
<style|generic>
<\body>
<section|MATH662 Homework 3, Issued 2/5/14, Due 2/13/14>
1. Complete the proof from notes in Lect08 to obtain the condition number
of computing root <math|j>
<TeXmacs|1.0.7.16>
<style|<tuple|generic|maxima>
<\body>
<section|MATH662 Lecture 3 (1/16/14)>
<subsection|Matrix norms>
<\remark>
The norm induced by the scalar product is widely used (Euclidean
dist
MATH662: Quiz 1
Answer the following questions in no more than 10 minutes.
., vn in a vector space (V , R, +) form a basis?
1. Do n linearly independent vectors v1,
Only if spancfw_v1, , vn = V
2. Is
MATH662 Homework 1, Issued 1/16/14, Due 1/23/14
1. Let entries of A Cm n , B Cn p depend on x. Prove:
d
dA
dB
(A(x)B (x) =
B+A
dx
dx
dx
Let m = n, A be nonsingular. Prove
d
dA 1
(A 1(x) = A1(x)
A (x)
MATH662: Quiz 2
Answer the following questions in no more than 10 minutes.
1. Can you interpret a vector u Cm as a function? If so, dene the three components of the function.
2. Are the functions cfw_
CHAPTER 11
Finite element methods
1. Preliminaries
For a number of applications the restrictions imposed by nite dierence or
spectral methods with respect to the computational grid are too severe. Thi