You are on page 1of 4

INSTITUTE OF SPACE TECHNOLOGY

MAT-721
Advanced Numerical Techniques
OHT-02-A
Fall 2015 December 22, 2015
Instructor: Faizan Ahmed Max Marks: 20

Question 1: (5 marks).
Starting from the initial guess(1,2), use Newton’s method to solve the system (perform three
iterations):
𝑓1: = 3𝑥12 − 𝑥22 = 0
𝑓2 ≔ 3𝑥1 𝑥22 − 𝑥13 − 1 = 0

The formula for Newton;s method is:


𝑋𝑖 = 𝑋𝑖−1 − 𝐽−1 (𝑋𝑖−1 )𝐹(𝑋𝑖−1 )
Here,
𝑭(𝑿) = [𝟑𝒙𝟐𝟏 − 𝒙𝟐𝟐
𝟑𝒙𝟏 𝒙𝟐𝟐 − 𝒙𝟑𝟏 − 𝟏]
And
𝝏𝒇𝟏 𝝏𝒇𝟏
𝝏𝒙𝟏 𝝏𝒙𝟐
𝑱=
𝝏𝒇𝟐 𝝏𝒇𝟐
[𝝏𝒙𝟏 𝝏𝒙𝟏 ]
𝟔𝒙𝟏 −𝟐𝒙𝟐
=[ 𝟐 𝟐 ]
𝟑𝒙𝟐 − 𝟑𝒙𝟏 𝟔𝒙𝟏 𝒙𝟐
And
𝟏 𝟔𝒙𝟏 𝒙𝟐 +𝟐𝒙𝟐
𝑱−𝟏 = 𝟐 𝟐
[
𝟐 −𝟑𝒙𝟐 + 𝟑𝒙𝟐
]
𝟑𝟔𝒙𝟏 𝒙𝟐 + 𝟐𝒙𝟐 (𝟑𝒙𝟐 − 𝟑𝒙𝟏 ) 𝟐 𝟏 𝟔𝒙𝟏
Now, you just need to keep pluging values and obtain answers.
Here
𝑿𝟎 = (𝟏, 𝟐)𝑻
So,
𝑿𝟏 = 𝑿𝟎 − 𝑱−𝟏 (𝑿𝟎 )𝑭(𝑿𝟎 )
𝟏 𝟏 𝟔(𝟏)(𝟐) +𝟐(𝟐) 𝟑(𝟏)𝟐 − (𝟐)𝟐
=[ ]− [ ] [ ]
𝟐 𝟑𝟔(𝟏)𝟐 (𝟐) + 𝟐(𝟐)(𝟑(𝟐)𝟐 − 𝟑(𝟏)𝟐 ) −𝟑(𝟐)𝟐 + 𝟑(𝟏)𝟐 𝟔(𝟏) 𝟑(𝟏)(𝟐)𝟐 − (𝟏)𝟑 − 𝟏
𝑹𝒆𝒔𝒕 𝒊𝒔 𝒄𝒐𝒎𝒑𝒖𝒕𝒂𝒕𝒊𝒐𝒏𝒔.

OR
Prove that a polynomial function of degree 𝑛 can have at most 𝑛 − 1 extremas over (−∞, ∞).
Sol:
The extreme values lies either on the critical point or on the point where the function is not
differentiable. Since the polynomial function is differentiable everywhere so the only candidates for
maximizers are critical points. Since the derivative of degree n polynomial is a polynomial of degree
𝑛 − 1. The 𝑛 − 1 degree polynomial can have at most 𝑛 − 1 roots according to fundamental theorem
of linear algebra.

Question 2(5 marks):


Find the first three iterations of the Power method applied to the following matrix
INSTITUTE OF SPACE TECHNOLOGY

1 −1 0
[−2 4 −2]
0 −1 1
Use 𝑥 0 = (1, −1,2)𝑡
Sol:
You just need to find 𝑥1 = 𝐴𝑥 0 , 𝑥 2 = 𝐴𝑥1 , 𝑥 3 = 𝐴𝑥 3 ,
OR
Find the extreme values of:
𝑓(𝑥, 𝑦, 𝑧) = 2𝑥𝑦 + 2𝑥𝑧 + 2𝑦𝑧 Subject to 𝑥 2 + 𝑦 2 + 𝑧 2 = 1
Sol:
The Lagrange function is:
𝐿(𝑥, 𝑦, 𝜆) = 𝑓(𝑥, 𝑦, 𝑧) + 𝜆(𝑥 2 + 𝑦 2 + 𝑧 2 − 1)
The critical points are solution of the system:
𝜕𝐿
= 2𝑦 + 2𝑧 + 2𝜆𝑥 = 0
𝜕𝑥
𝜕𝐿
= 2𝑥 + 2𝑧 + 2𝜆𝑦 = 0
𝜕𝑦
𝜕𝐿
= 2𝑥 + 2𝑦 + 2𝜆𝑧 = 0
𝜕𝑧
𝜕𝐿
= 𝑥2 + 𝑦2 + 𝑧2 − 1 = 0
𝜕𝜆

Solve the above system to find critical points and then use Hessian matrix to determine if they are
point of maxima or minima,
Question 3: (4 marks).
Determine the singular value decomposition of the matrix:
2 1
( )
0 1
Sol:
4 2
Find the eigenvalues of the matrix𝐴𝑇 𝐴 = [ ], the eigenvalues are:
2 2
.
So S matrix is :

You can alsofind the eigenvalues to find matrix V.

And now use it to find the matrix U.

Question 4: a) (3 marks). Use Euler method to approximate the solution to the initial value problem:
𝑦 ′ = 𝑐𝑜𝑠 2𝑡 + 𝑠𝑖𝑛 3𝑡, 0 ≤ 𝑡 ≤ 1, 𝑦(0) = 1, 𝑤𝑖𝑡ℎ ℎ = 0.25
Sol:

Its just computations nothing special.

b)(3 marks) Show that the following initial value problems have unique solution
INSTITUTE OF SPACE TECHNOLOGY

𝑦 ′ = 𝑦𝑐𝑜𝑠 𝑡, 0 ≤ 𝑡 ≤ 1, 𝑦(0) = 1
Sol: You need to check if the function 𝒇(𝒕, 𝒚) = 𝒚𝒄𝒐𝒔 𝒕 is:
1. Is continuous
2. Satisfy Lipschitz conditions in the variable y.
INSTITUTE OF SPACE TECHNOLOGY

MAT-721
Advanced Numerical Techniques
OHT-02-B
Fall 2015 December 22, 2015
Instructor: Faizan Ahmed Max Marks: 20

Question 1: (5 marks).
Use Newton’s method to solve the following system (perform three iterations with initial guess (1,2)).
𝑥12 + 𝑥22 − 𝑥1 = 0
𝑥12 − 𝑥22 − 𝑥2 = 0

OR
Determine the value of 𝑎 such that the matrix:
1 𝑎
( )
𝑎 4
Is: a) positive definite b) negative definite.

Question 2(5 marks):


Find the first three iterations of the Power method applied to the following matrix
1 −1 0
[−2 4 −2]
0 −1 2
Use 𝑥 0 = (−1,2,1)𝑡
OR
Find extreme value of:
𝑓(𝑥, 𝑦, 𝑧) = 3𝑥 2 + 2𝑦 2 + 3𝑧 2 2𝑥𝑧 subject to 𝑥 2 + 𝑦 2 + 𝑧 2 = 1

Question 3: (4 marks).
Determine the singular value decomposition of the matrix:
−1 1
( )
1 1
Question 4: a) (3 marks). Use Euler method to approximate the solution to the initial value problem:
𝑦 ′ = 1 + (𝑡 − 𝑦)^2, 0 ≤ 𝑡 ≤ 3, 𝑦(2) = 1, 𝑤𝑖𝑡ℎ ℎ = 0.25
b)(3 marks) Show that the following initial value problems have unique solution
2
𝑦 ′ = ( ) 𝑦 + 𝑡 2 𝑒 𝑡 , 0 ≤ 𝑡 ≤ 2, 𝑦(1) = 0
𝑡

You might also like