AMSC 660 / CMSC 660 Scientific Computing I (Section 0101)

Information for Fall 2004


Dianne P. O'Leary


Basic Information from Fall 2004:
  • Note that some links are broken.
  • Course Information and Syllabus: postscript and pdf
  • References
  • UMCP Code of Academic Integrity
  • Information about the computer accounts
  • Survival Guide for Scientific Computing

    • Lecture Notes
  • CMSC/AMSC 460 notes (for background)
  • Notes for Introduction
  • Error Analysis postscript or pdf (The pdf file is fuzzy, so use the postscript if at all possible. Later notes are Latex and the pdf will be fine.)
  • Notes for Monte Carlo Methods Unit
  • pdf (Reposted 09/23/04)
  • Demonstration of generating random numbers: demorand.m
  • Demonstration of Monte Carlo integration: demomcint.m
  • Demonstration of Monte-Carlo counting:
    The KRS algorithm for the monomer/dimer covering problem: krs.m
  • Notes for Matrix Factorizations Unit
  • pdf
  • A homework assignment (with solution) from 2002: pdf . (Acknowledgement: This problem is a variant of one by James Nagy.)
  • The data: Hmwk1Data.mat .
    If this .mat file displays in your window, then it won't download properly. If you have this trouble in Netscape, see if right-clicking on the link brings up a download option for you.
  • The demonstration program: Hmwk1demo.m .
  • The documentation model: gmres.m
  • Answers: pdf
  • Matlab code that solves the problem
  • Original image: origimage.mat
  • Notes for Optimization Unit
  • pdf (reposted 10/28, with a few corrections and a new slide on conjugate gradients)
  • Notes on the (linear) conjugate gradient method pdf
  • Notes for Ordinary Differential Equations Unit
  • Review of 460 material: postscript or pdf (Reposted 11/09 to fix stability typos)
  • More advanced material
  • Notes for Nonlinear Equations Unit
  • pdf (reposted 11/23, with one new slide on p.10)
  • A case study (pdf)
  • Final Case Study: Epidemiology
  • Discrete model
  • Continuous model and answers for Discrete model
  • Answers for continuous model
  • Homework 1: Due Sept 14.
  • Frequently asked questions.
  • Solution (pdf)
  • problem2.m
  • dist_to_ellipse.m
  • find_dist.m
  • pitfalls.m
  • Homework 2:
  • Due Oct 5.
  • Frequently asked questions.
  • pdf
  • problem1.m
  • problem2.m
  • problem3.m
  • problem4.m
  • quasirand.m
  • problem5.m (Actually, it solves problem 6)
  • quadmcqr.m
  • partition.m
  • Homework 3:
  • Due October 26 (pdf)
  • spectdata2.mat containing K, g, nbins, bin centers, and energy ranges for each bin. If this .mat file displays in your window, then it won't download properly. If you have this trouble in Netscape, see if right-clicking on the link brings up a download option for you.
  • Frequently asked questions.
  • Grading: As you know, the homework was listed as being worth 35 points. You have 2 options:
  • The default: I will take the score you got, divide by 40, and multiply by 35.
  • If you request it: I will compute your score by dividing your score on Problem 1 by 2 and adding it to your score on the other problems.
    • Solution (pdf) (Reposted 11/04/04 with more detail at the top of p.3)
  • problems1_and_3.m (reposted 11/23)
  • makedata.m
  • Homework 4:
  • Due November 23 (pdf) Use the data from Homework 3.
  • Frequently asked questions.
  • Solution (pdf)
  • problem4.m
  • stls.m
  • Homework 5:
  • Due December 9 (pdf)
  • Frequently asked questions.
  • The Latex file that produces hmwk5.pdf. You are welcome to make use of this as a template if you write your project in Latex.
  • New! Solution (pdf)
  • New! problem1.m
  • New! osc1.m
  • New! osc2.m
  • Term Project:
  • Description
  • Claimed topics and further information

    • Quizzes:
  • quiz 1: Sep 9 Questions and Answers (reposted Sep 13 09:34) .
  • quiz 2: Sep 21 Hints. (posted noon on Sept 16)
    Questions and Answers .
  • quiz 3: Sep 30 Hints. (posted 12:20 on Sept 28)
    Questions and Answers .
  • quiz 4: Oct 12 Hints. (posted 11:20 on Oct 7)
    Questions and Answers .
  • quiz 5: Oct 21 Hints. (posted 11:55 on Oct 19)
    Questions and Answers .
  • quiz 6: Nov 2 Hints. (posted 11:55 on Oct 28)
    Also see some supplementary material (in green) in the lecture notes. If you didn't catch in class the definition of an exact linesearch: In an exact linesearch we find the value of alpha that exactly minimizes f(x + alpha p). We can do this for quadratic functions, since in that case a formula for alpha can be derived, but in general exact linesearch is impossible and is only interesting because a lot of theorems demand it. Questions and Answers .
  • quiz 7: Nov 11 Hints. (posted 12:03 on Nov 9)
    Questions and Answers .
  • quiz 8: Nov 23 Hints. (posted 11:45 on Nov 18)
    Questions and Answers .
  • quiz 9: Dec 7 Hints. (posted 12:15 on Dec 2)
    New! Questions and Answers .