Scribd
Upload
9 views

Numeric-Exam

This document is an examination paper for the Advanced Numerical Methods and Computer Software course at Arba Minch University, dated February 5, 2021. It consists of six problems covering topics such as sequence convergence, system of equations, quadratic minimization, optimization methods, and algorithm implementation. Students are instructed to attempt all problems without using unauthorized materials and to write clearly in the provided spaces.

Uploaded by

emru erade
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views

Numeric-Exam

This document is an examination paper for the Advanced Numerical Methods and Computer Software course at Arba Minch University, dated February 5, 2021. It consists of six problems covering topics such as sequence convergence, system of equations, quadratic minimization, optimization methods, and algorithm implementation. Students are instructed to attempt all problems without using unauthorized materials and to write clearly in the provided spaces.

Uploaded by

emru erade
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

Department of Mathematics

College of Natural Science


Arba Minch University
Advanced Numerical Methods and Computer Software (Math 823)
2013 EC

Date: February 05, 2021

Time: 3 hours and half

Instructor: Surafel L Tilahun (PhD)

Name: _____________________ Student number: ___________________

Email: _____________________

Instructions:

- This examination contains 6 problems, attempt all the problems


- Using any additional notes and materials other than those provided by the invigilators is not
allowed
- Write neatly and clearly in the space provided

For Instructor’s use only

Q1 Q2 Q3 Q4 Q5 Q6
4pts 6pts 15pts 10pts 14pts 11pts

Good luck

1
Name: _____________________________ Student number: _______________________

1. Consider a sequence of numbers given by {𝑎 } where 𝑎 = . Discuss the speed of


convergence of the sequence. [4pts]

2
Name: _____________________________ Student number: _______________________

2. Find an approximate solution for the system of equations given below:


𝑥 − 2𝑥 + 𝑥 = 1
𝑥 𝑒 +𝑥 =5
𝑥 +𝑥 −𝑥 =0
[6pts]

3
Name: _____________________________ Student number: _______________________

3. Consider the quadratic minimization problem given by:

min 𝑓(𝑥)
∈ℝ

3 −2 4
where 𝑓(𝑥) = 𝑥 𝐴𝑥 − 𝑏 𝑥 + 𝑐 for 𝐴 = , 𝑏= and 𝑐 = 2
−2 6 2
a) Show that 𝐴 is positive definite [2pts]
b) Determine ∇𝑓(𝑥) and 𝐻(𝑥) [2pts]
c) Find the local minima using the necessary and sufficient second order conditions. Which
one the is global minima [4pts]

4
Name: _____________________________ Student number: _______________________

d) Determine the exact line search parameter 𝑡 which minimize the 𝑓(𝑥 − 𝑡∇𝑓(𝑥 ))
where the starting point 𝑥 = (0, 0) [4pts]
e) Calculate one step of the steepest descent method from 𝑥 = (0, 0) using the step
length computed in (d) [3pts]

5
Name: _____________________________ Student number: _______________________

4. Answer the following based on Newton Method and Broyden Method.


a. Compare how the two algorithms work [5pts]
b. Discuss the weakness and strength of the two methods [5pts]

6
Name: _____________________________ Student number: _______________________

5. Consider a minimization problem given below.


min 𝑓(𝑥) = 𝑥 + 2𝑥 𝑥 + 2𝑥 + 𝑥
∈ℝ

Based on the initial point of 𝑥 ( ) = (−3, 4) and radius of the trust region at the start of the
iteration being Δ = 1, show two iterations of solving the given problem using

a. The Cauchy point method [7pts]

7
Name: _____________________________ Student number: _______________________

b. The dogleg method [7pts]

8
Name: _____________________________ Student number: _______________________

6. Consider an optimization problem given below:

max 𝑓(𝑥) = −20 + (10 cos 2𝜋𝑥 − 𝑥 )


. .

−10 ≤ 𝑥 ≤ 10
𝑥∈ℝ
Suppose a pseudo code to solve the given problem is also given as follows:

Input: Objective function (f(x)) and feasible region


Algorithm parameters: 𝜆, 𝑁, 𝑀𝑎𝑥𝑖𝑚𝑢𝑚 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑖𝑡𝑒𝑟𝑎𝑡𝑖𝑜𝑛
Randomly generate N feasible solutions, 𝑥 ′𝑠
Compute the performance of each of these solutions, 𝑓(𝑥 )′𝑠
For all the solutions update the solutions using
𝑥 : = 𝑥 + 𝜆𝑟, where 𝑟 is a random unit direction
If the updated solution is feasible and is better than the old one
Keep the new solution otherwise keep the old solution
Terminate if the maximum number of iteration is reached.
Output: the best solution found

a. Write a MATLAB code of the algorithm to solve the given optimization problem. [7pts]

9
Name: _____________________________ Student number: _______________________

b. Discuss the effects of the algorithm parameters [4pts]

10

You might also like