Computational Methods and Software - Final Exam Problems | CS 417, Exams of Computer Science

Material Type: Exam; Professor: Morris; Class: Computational Methods and Software; Subject: Computer Science; University: Old Dominion University; Term: Spring 2010;

Typology: Exams

Pre 2010

Uploaded on 05/07/2010

timpkmn89
timpkmn89 🇺🇸

1 document

1 / 4

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
CS 417
You have 1 week to complete this exam. Use the online Submit for your completed
solutions. If you scan your documents please make sure that it is a legible .pdf file.
Answer each question as completely as possible. Be clear and succinct. Quality of
response is more important than quantity. Take time to "THINK", they may not be as
hard as they originally appear. It is open book and open notes; the only restriction is that
you are not allowed to talk to any other person about this exam. You may use the library
or any computer lab as well. Any violation of this rule is a violation of the Old Dominion
University Honor Code.
1. A researcher is performing a series of experiments and has accumulated the following
data. He theorizes that the physics dictates a solution of the form
Do a best fit analysis of this data in the manner of your choosing.
The data are presented in this table:
Input Output
-3.0226 56.7484
7.9782 296.324
0.9814 3.87153
0.035 1.90113
-7.9882 345.021
-5.0286 143.52
5.0298 113.405
7.9909 297.3
0.0287 1.91802
6.973 224.195
2. Consider LU decomposition on very, very large matrices. It is a direct method when
applied in the standard manner, e.g. O(N3). Can you come up with a method to reduce the
computational complexity of LU for very, very large SPARSE matrices? Describe your
solution OR tell me why it is not possible.
3. Write out the stencil for the following equation, use only CENTERED
approximations.
pf3
pf4

Partial preview of the text

Download Computational Methods and Software - Final Exam Problems | CS 417 and more Exams Computer Science in PDF only on Docsity!

CS 417

You have 1 week to complete this exam. Use the online Submit for your completed solutions. If you scan your documents please make sure that it is a legible .pdf file. Answer each question as completely as possible. Be clear and succinct. Quality of response is more important than quantity. Take time to "THINK", they may not be as hard as they originally appear. It is open book and open notes; the only restriction is that you are not allowed to talk to any other person about this exam. You may use the library or any computer lab as well. Any violation of this rule is a violation of the Old Dominion University Honor Code.

  1. A researcher is performing a series of experiments and has accumulated the following data. He theorizes that the physics dictates a solution of the form Do a best fit analysis of this data in the manner of your choosing. The data are presented in this table: Input Output -3.0226 56. 7.9782 296. 0.9814 3. 0.035 1. -7.9882 345. -5.0286 143. 5.0298 113. 7.9909 297. 0.0287 1. 6.973 224.
  2. Consider LU decomposition on very, very large matrices. It is a direct method when applied in the standard manner, e.g. O(N^3 ). Can you come up with a method to reduce the computational complexity of LU for very, very large SPARSE matrices? Describe your solution OR tell me why it is not possible.
  3. Write out the stencil for the following equation, use only CENTERED approximations.
  1. Generate the matrix equation, Ax=b, that represents the following differential equation on a Unit Square. The grid is regular in each direction and has 10 nodes along the x-axis, 3 along the y-axis. The boundary conditions are dirchlet and have a value of –4. Just write out enough of the matrix to illustrate any pattern that exists. (n.b. nx does not equal ny)

f  

  1. Given the Laplace equation in 2-Dimensions on a unit square, how would you handle the situation where the step size starts at 0, 0 = .5 in both x and y directions but it then halved with each additional step? How would it change the stencil? How would it change the matrix? The right hand side?
  1. The Schrödinger equation is an equation that describes how the quantum state of some physical system changes in time. This equation is as central to quantum mechanics as Newton’s laws are to classical mechanics. The standing wave version of this equation is time independent and is stated as: Where: E is the total energy of a system. V is the potential energy of a particle, is the amplitude at position r, where r (x,y)
  • is the kinetic energy (m is the mass, ћ is Plank’s constant) Assuming ћ, m, V and E are constant write out the finite difference approximation used to solve this equation for Ψ(r).

on a square domain are there any special boundary conditions that we will have to deal with?