Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Q1.

Prove that the induced norm


∥ 𝐴 ∥2 = 𝜎1 [the largest singular value of A]
Assume A to be a rectangular matrix of dimension (m x n)
Q2. For any inner product of two vectors prove the Cauchy-Schwartz inequality

|⟨𝑥, 𝑦⟩| ≤ √⟨𝑥. 𝑥⟩ . ⟨𝑦. 𝑦⟩


Q3. Find the induced 1-norm of the matrix
1 2 3
[3 2 1]
1 1 1
𝑥 𝑇 𝐒𝑥
Q4. Find the maximum value of
𝑥𝑇𝑥
When S is symmetric and positive definite.
Q5. For a symmetric matrix A Prove
1
∥ 𝐴 ∥𝐹 = [∑ 𝜎12 ]2
Where 𝜎1 , 𝜎2 , … … 𝜎𝑁 are the singular values of 𝐀
Q6. For the following system the input output relation is given as
𝑑2 𝑥(𝑡)
𝑑𝑡2
+ 𝑑𝑥(𝑡)
𝑑𝑡
+ 𝑥 (𝑡 ) = 𝑢 (𝑡 )

u(t) x(t)
h(t)

The impulse response x(t) is sampled at 𝑇𝑠 = 1 sec and the samples of x[n] are
arranged as
𝑥0 𝑥10 … 𝑥1000
𝑥1 𝑥11 … 𝑥1001
X=[ ⋮ ⋮ ⋮ ⋮ ]
𝑥9 𝑥19 ⋯ 𝑥1019
Find the most compact orthogonal representation of the vector space X. Show the
method.
Q7. Given 𝐴 = [1 0]
2 1
An iterative formula is given as
1
𝑋𝑘+1 = (𝑋𝑘 + 𝐴𝑋𝑘−1 )
2
1 0.1
With 𝑋0 = [ ]
0.1 1
Find Xk, for when k is large
Q8. Given the following matrix
−2 1 1 0 0 ⋯ 0
0 −2 1 1 . ⋯ 0
H= 0 0 −2 1. 1 ⋯ 0
⋮ ∙ . . . ⋮
[ 0 0 0 ⋯ −2 1 1 ]14×16

Find the closest approximation of the induced nor ∥ H ∥2

Q9. Express
1 1 2
𝐴 = [ 2 3 1] = 𝑄𝐿
3 2 3
Where 𝑄 𝑇 𝑄=1, L is lower triangular.

Q10. If 𝑝−1 + 𝑞−1 = 1, and p,q >1, Prove, that


|𝑣 𝑇 𝑤| ≤∥ 𝑣 ∥𝑝 ∥ 𝑤 ∥𝑞
Q11. Four eigenvectors of a permutation matrix P are 𝑥1 = (1, 1, 1, 1),𝑥2 =
(1, 𝑖, 𝑖 2 , 𝑖 3 ), 𝑥3=(1,𝑖 2 , 𝑖 4 , 𝑖 6 ), and 𝑥4= (1,𝑖 3 , 𝑖 6 , 𝑖 9 ). The eigenvectors are the columns
of the 4 by 4 Fourier matrix F.
1 1 1 1
𝐅 1 1 𝑖 −1 −𝑖 ] 𝑻
Show that 𝑄 = = [ has orthonormal columns : 𝑸 𝑸 = 𝑰
2 2 1 𝑖2 1 −1
1 𝑖3 −1 𝑖

Q12. Haar wavelets are orthogonal vectors (columns of W) using only 1, -1, and 0.
For 𝑛 = 4
1 1 1 0
Find 𝑊 𝑇 W and 𝑊 −1 and the
𝑊 = [1 1 −1 0]
1 −1 0 1 eight Haar wavelets for n = 8.
1 −1 0 −1

Q13. In the Cholesky factorization S = A A, with A = D L , the square roots of the


T 1/2 T

pivots are on the diagonal of A. Find A (upper triangular) for


9 0 0 1 1 1
𝑆 = [0 1 2] and 𝑆 = [1 2 2] .
0 2 8 1 2 7

1
Q14. For 𝐹1 (𝑥, 𝑦) = 4 𝑥 4 + 𝑥 2 𝑦 + 𝑦 2 and 𝐹2 (𝑥, 𝑦) = 𝑥 3 + 𝑥𝑦 − 𝑥 find the
second derivative matrices 𝐻1 and 𝐻2 (the Hessian matrices).
𝜕 2 𝐹 ⁄𝜕𝑥 2 𝜕 2 𝐹 ⁄𝜕𝑥𝜕𝑦
Test for minimum 𝐻 = [ 2 ] is positive definite
𝜕 𝐹 ⁄𝜕𝑦 𝜕 𝑥 𝜕 2 𝐹 ⁄𝜕𝑦 2
𝐻1 is positive definite so 𝐹1 is concave up (= convex). Find the minimum point of
𝐹1 .Find the saddle point of 𝐹2 .
Q15. Let A ∈ 𝑀𝑛 be given. (a) Show that the sum of the entries in each row of A is
1 if and only if 1 ∈ 𝜎(𝐴) and the vetor 𝑒 = [1,1, … , 1]𝑇 is an associated
eigenvector,that is,𝐴𝑒 = 𝑒.(b) Suppose that the sum of the entries in each row of
A is 1. If A is nonsingular,show that the sum of the entries in each row of A -1 is also
1.Moreover, for any given polynomial p(t), show that the sums of the entries in
each row p(A) are equal.Equal to what?

Q16. Let A ∈ 𝑀𝑛 and let 𝜆, 𝑥 be an eigenvalue-eigenvector pair of A. Show that 𝑥


is an eigenvector of adj A

Q17. Let 𝑥, 𝑦 ∈ 𝑪𝑛, 𝑎 ∈ 𝑪, and 𝐵 ∈ 𝑀𝑛. Consider the bordered matrix 𝐴 = [𝑦𝐵∗ 𝑥
]∈
𝑎
𝑀𝑛+1.(a)Use(0.8.5.10)to Show that

𝑃𝐴 (𝑡) = (𝑡 − 𝑎)𝑃𝐵 (𝑡) − 𝑦 ∗ (𝑎𝑑𝑗(𝑡𝐼 − 𝐵))𝑥

(b) If B = 𝜆𝐼𝑛, deduce that


𝑃𝐴 (𝑡) = (𝑡 − 𝜆)𝑛−1 (𝑡 2 − (𝑎 + 𝜆)𝑡 + 𝑎𝜆 − 𝑦 ∗ 𝑥)

of [𝜆𝐼 𝑥
and conclude that the eigenvalues 𝑛
] are 𝜆 with multiplicity n - 1, together
𝑦∗ 𝑎
with 12 (𝑎 + 𝜆 ± ((𝑎 − 𝜆)2 + 4𝑦 ∗𝑥)1⁄2 ).

Q18. Suppose that all the entries of 𝐴 = [𝑎𝑖𝑗 ] ∈ 𝑀𝑛 are either zero or one, and
suppose that all the eigenvalues 𝜆1 , … , 𝜆𝑛 of A are positive real numbers.Explain
why det A is a positive integer, and provide details for the following:
1
𝑛 ≥ 𝑡𝑟 𝐴 = 𝑛 (𝜆1 + ⋯ + 𝜆𝑛 )𝑛 ≥ 𝑛(𝜆1 … 𝜆𝑛 ) 1⁄𝑛

= 𝑛(det 𝐴)1⁄𝑛 ≥ 𝑛
Conclude that all 𝜆𝑖 = 1, all 𝑎𝑖𝑖 = 1, and det A =1.

Q19. Prove Cauchy’s identity det(𝐴 + 𝑥𝑦 𝑇 ) = det 𝐴 + 𝑦 𝑇 (𝑎𝑑𝑗 𝐴)𝑥, valid for
any 𝐴 ∈ 𝑀𝑛 and 𝑥, 𝑦 ∈ 𝑪𝑛
Q20. Let 𝑥, 𝑦 ∈ 𝑪𝑛 be given and suppose that 𝑦 ∗𝑥 ≠ −1.(a) Verify that(𝐼 + 𝑥𝑦 ∗)−1 =
𝐼 − 𝑐𝑥𝑦 ∗ , in which 𝑐 = (1 + 𝑦 ∗ 𝑥)−1.(b) Let ⋀ = 𝑑𝑖𝑎𝑔(𝜆1 , … , 𝜆𝑛) and suppose that
𝑦 ∗ 𝑥 = 0. Explain why the eigenvalues of

𝐴 = (𝐼 + 𝑥𝑦 ∗ ) ∧ (𝐼 − 𝑥𝑦 ∗ ) =∧ +𝑥𝑦 ∗ ∧ − ∧ 𝑥𝑦 ∗ − (𝑦 ∗ ∧ 𝑥)𝑥𝑦 ∗

are 𝜆1 , … , 𝜆𝑛 .Notice that A has integer entries if the entries of x,y, and ∧ are
integers. Use this observation to construct an interesting 3-by-3 matrix with
integer entries and eigenvalues 1,2, and 7; verify that your construction has the
asserted eigenvalues.

Q21. Brauer’s theorem. Let 𝑥, 𝑦 ∈ 𝑪𝑛 , 𝑥 ≠ 0, and 𝐴 ∈ 𝑀𝑛 . Suppose that 𝐴𝑥 =


𝜆𝑥 and let the eigenvalues of A be 𝜆, 𝜆2 , … , 𝜆𝑛. What are the eigenvalues of 𝐴 +
𝑥𝑦 ∗ ? First observe that (𝑡 − 𝜆)𝑥 = (𝑡𝐼 − 𝐴)𝑥 implies that (𝑡 − 𝜆)
adj(𝑡𝐼 − 𝐴)𝑥 = adj (𝑡𝐼 − 𝐴)(𝑡𝐼 − 𝐴)𝑥 = det(𝑡𝐼 − 𝐴)𝑥,that is,
Q22. Solve this matrix problem basic to Fisher’s Linear Discriminant
Analysis:
𝑥 𝑇 𝑺𝑥
If 𝑅 (𝑥) = and 𝑆 = 𝑢𝑢 𝑇 what vector 𝑥 minimizes R(𝑥) ?
𝑥 𝑇 𝑴𝑥

You might also like