Matrix Methods Exam Solutions: Orthogonal Basis, Gram Matrix, Least Squares, Exams of Mathematics

Solutions to exam #2 of the appm 3310: matrix methods course, covering topics such as finding an orthogonal basis for the range of a matrix, determining the dimension of the kernel, finding a basis for the coker, calculating the gram matrix, and finding the least squares solution to a linear system.

Typology: Exams

2012/2013

Uploaded on 02/23/2013

shuja
shuja 🇮🇳

4.3

(6)

51 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
APPM 3310: Matrix Methods Exam #2 Solutions April 2, 2010
Problem 1. (25 points) Suppose we have A=
1 0 2
1 1 1
0 1 1
.
(a) (9 pts) Find an orthogonal basis for range(A).
(b) (8 pts) What is the dimension of the kernel of A?
(c) (8 pts) Find a basis for coker(A).
Solution:
(a) A=
1 0 2
1 1 1
0 1 1
102
011
011
102
011
000
.So range(A) =
span
1
1
0
,
0
1
1
. Now using Gram Schmidt, starting with ~v1=
1
1
0
, we have
~v2=~w2h~w2, ~v1i
k~v1k2~v1=
0
1
1
1
2
1
1
0
=
1/2
1/2
1
so an orthogonal basis of range(A) is
1
1
0
,
1/2
1/2
1
.
(b) By the Fundamental Theorem of Linear Algebra dim(ker A) = mrank(A)=32 = 1
(c) Solve ATy= 0,
1 1 0 0
0 1 1 0
21 1 0
1 1 0 0
0 1 1 0
0 1 1 0
1 1 0 0
0 1 1 0
0 0 0 0
x+y= 0
y+z= 0
z=z
x=z
y=z
z=z
thus coker(A) = span
1
1
1
.
Problem 2. (25 points)
(a) (15 pts) Find the Gram matrix generated by the vectors p1(x) = xand p2(x) = x2with
respect to the L2inner product on [0,1]. Show all work.
(b) (10 pts) Is the Gram matrix described in part (a) positive definite? Why or why not?
Justify your answer.
Solution:
(a) Note hx, xi=R1
0x2dx =x3
3
1
0= 1/3 and hx, x2i=R1
0x3dx =x4
4
1
0= 1/4 and hx2, x2i=
R1
0x4dx =x5
5
1
0= 1/5 so
K=hx, xi hx, x2i
hx2, xi hx2, x2i=1/3 1/4
1/4 1/5
pf2

Partial preview of the text

Download Matrix Methods Exam Solutions: Orthogonal Basis, Gram Matrix, Least Squares and more Exams Mathematics in PDF only on Docsity!

APPM 3310: Matrix Methods — Exam #2 Solutions — April 2, 2010

Problem 1. (25 points) Suppose we have A =

(a) (9 pts) Find an orthogonal basis for range(A). (b) (8 pts) What is the dimension of the kernel of A? (c) (8 pts) Find a basis for coker(A).

Solution:

(a) A =

 (^). So range(A) =

span

. Now using Gram Schmidt, starting with ~v 1 =

, we have

~v 2 = w~ 2 −

〈 w~ 2 , ~v 1 〉 ‖~v 1 ‖^2

~v 1 =

so an orthogonal basis of range(A) is

(b) By the Fundamental Theorem of Linear Algebra dim(ker A) = m − rank(A) = 3 − 2 = 1

(c) Solve AT^ y = 0, 

−x + y = 0 y + z = 0 z = z

x = −z y = −z z = z

thus coker(A) = span

Problem 2. (25 points)

(a) (15 pts) Find the Gram matrix generated by the vectors p 1 (x) = x and p 2 (x) = x^2 with respect to the L^2 inner product on [0, 1]. Show all work.

(b) (10 pts) Is the Gram matrix described in part (a) positive definite? Why or why not? Justify your answer.

Solution:

(a) Note 〈x, x〉 =

0 x

(^2) dx = x^3 3

∣^1

0 = 1/3 and^ 〈x, x

2 〉 = ∫^1

0 x

(^3) dx = x^4 4

∣^1

0 = 1/4 and^ 〈x

(^2) , x (^2) 〉 = ∫ (^1) 0 x

(^4) dx = x^5 5

∣^1

0 = 1/5 so

K =

〈x, x〉 〈x, x^2 〉 〈x^2 , x〉 〈x^2 , x^2 〉

(b) The Gram matrix in part (a) is positive definite since x and x^2 are linearly independent.

Problem 3. (20 points) Find the least squares solution to the linear system:   

x y

Solution: Here K = AT^ A =

and f = AT^ b =

. Now solve Kx = f , ( 2 2 3 2 7 6

→ 2 x^ + 2y^ =^3 5 y = 3

→ x^ =^9 /^10 y = 3 / 5

so the least squares solution is x∗^ =

Problem 4. (20 points) Give a brief answer to each question. Show all work.

(a) Consider the norms ‖ · ‖ 3 and ‖ · ‖ 5 on Rn. Now define the F-norm, ‖ · ‖F , as

‖x‖F =

‖x‖ 5 + ‖x‖ 3

, for any x ∈ Rn.

Show that the F-norm defined above satisfies the Triangle inequality. Show all work.

(b) Does the expression 〈u, v〉 = u 1 v 1 + 2u 1 v 2 define an inner product on R^2? Why or why not? Explicitly justify your answer.

Solution:

(a) Show ‖x + y‖F ≤ ‖x‖F + ‖y‖F , so note

‖x + y‖F =

‖x + y‖ 5 + ‖x + y‖ 3

‖x‖ 5 + ‖y‖ 5 + ‖x‖ 3 + ‖y‖ 3

‖x‖ 5 + ‖x‖ 3

‖y‖ 5 + ‖y‖ 3

= ‖x‖F + ‖y‖F

where the inequality follows from the fact that both ‖ · ‖ 3 and ‖ · ‖ 5 satisfy the Triangle inequality.

(b) No, consider v =

then 〈v, v〉 = 1 · 1 + 2 · 1 · (−1) = − 1 < 0 which violates positivity.

Problem 5. (10 points) Does the set S =

z z

∣ z^ ∈^ C

form a vector subspace of C^2? Why or

why not? Justify your answer.

Solution: No, the set S is not a subspace since closure under scalar multiplication fails. Note, in general, for a complex scalar α ∈ C we have,

α ·

z z

αz αz

αz αz

, which implies α ·

z z

∈ / S.