Overview
- Instructor
- Jeffrey Wong ( main site)
- Email
- jtwong at math dot duke dot edu
- Office
- Physics 029B
- Office Hours
- M11-12, F4-5 or by appointment
- Lectures
- WF 3:05-4:20 (Physics 227)
- Syllabus
- Available here.
- Textbook
- Uri Ascher & Chen Greif, A First Couse in Numerical Methods and
excerpts from Cleve Moler’s book Numerical Methods in MATLAB (free online).
Description
Development of numerical techniques for accurate, efficient
solution of problems in science, engineering, and mathematics through the use of
computers. Linear systems, nonlinear equations, optimization, numerical
integration, differential equations, error analysis.
Prerequisites
solid understanding of fundamental concepts from linear
algebra is essential, including linearity, solving linear systems, eigenvalues and
eigenvectors. A course in multi-variable calculus (e.g. Math 212) is also required.
Experience with ordinary differential equations is recommended, but not necessary.
Grading
Grades will be assigned based on an assessment of your performance on homework, midterm exams, and a
final project. The components will be weighted (roughly) as follows:
- Weekly homework (20%), lowest dropped
- Two midterm exams (20% each)
- Final project (30%)
- There is no final exam for this course
Note that because this baseline score does not correspond directly to a letter grade; in a vacuum,
it does not provide much more information than a sense of how compontents are weighted. The interpretation
of all scores and course (letter) grade will depend on the final project and an assessment of your performance
as a whole.
Homework
- Homework will be assigned (roughly) weekly and will typically be due the
following Wednesday. Consult the schedule for due dates.
- No late homework will be accepted,
barring exceptional circumstances as per Duke policy.
- Working and studying in groups is encouraged
(you will get much more out of doing homework if you discuss it with others!).
However, you should write your own solutions to each problem in your own words.
- Solutions should be complete arguments; the process by which you arrive at
the solution is far more important than a correct answer. Assertions should be supported by computed data and code when it is needed.
- Homework pages must be stapled together with clearly readable work.
Solutions should be in the same order as in the list of assigned problems.
Computational work
- Some homework problems will require writing and running code. The official
choice of language for this course is Matlab; most examples/solutions will also
include python code. You may write your code in Matlab or python (for
python, use the numpy package).
- Collaboration is encouraged but the code you submit should be your own,
which includes not copy-pasting code from other sources. Avoid looking up
code online because it is difficult to un-see it when writing your own.
- Expectations for computational problems are detailed in the Guidelines for
computational problems document (on Piazza).
Ethics
Students are expected to follow the Duke Community Standard.
If a student is found responsible for academic dishonesty
through the Office of Student Conduct, the student will receive a
core of zero for that assignment. If a student’s admitted academic
shonesty is resolved directly through a faculty-student resolution
agreement approved by the Office of Student Conduct, the terms of that
ent will dictate the grading response to the assignment at issue.