Spring Semester 2001
Schedule: 6:15 - 7:45pm, Mon, Wed.
Location: South Physics Computer Lab (SP 205)
Instructor: Ben Bromley
Office: INSCC 218, Hours:TBA
Office: INSCC 2nd Floor, Hours:TBA
This course is a survey of modern numerical methods with programming exercises on Unix workstations. Topics include root finding, solving linear systems by direct and iterative methods, eigenvalue problems, interpolation and extrapolation, differentiation and integration, solution of ordinary and partial differential equations, elementary statistics, linear and nonlinear optimization, Fourier transforms.
Prerequisites: MATH 3150 and 3160 and either PHYCS 3730/5720 or CPSC 3200. A working knowledge of C++ and Maple is required.
The course will meet twice per week, Monday and Wednesdays, from 6:15pm to 7:45pm in the South Physics Computer Lab (2nd Floor). Instructor and TA office hours are given above.
Homework assignments will be posted on Fridays and are to be completed in 10 days, due midnight on the 2nd Monday after posting. These assignments will constitute 70% of the students' grades. Each student will be required to give a short presentation in class; participation in this way will be the basis for 10% of the grade. The remaining 20% will come from a final project and/or a final exam.
In addition to graded material, students will be expected to work on laboratory exercises in the South Physics Computer Lab. These will be posted weekly on Wednesdays, and students will be given time during class to complete them. No grades will be assigned for these exercises, however, their subject matter is strongly correlated with the homework....
There is no required textbook for this course, however, students are encouraged to purchase either Numerical Recipes in C by Press et al. (1992) (order from Cambridge University Press following this link -- but you may wish to hold off until a version in C++ is released???) or Numerical Analysis by Burden and Faires (available from amazon.com). Most of the homework assignemts and lecture material will be drawn from these two books.
The course will cover numerical algorithms that have proven to be generally useful in scientific computing. The intent is to give students experience with these algorthims so that they may select effective computational tools for problem solving. The algorithms considered here include those listed in the course description above. We will often introduce these algorithms in the context of specific physics problems--students are encouraged to make suggestions regarding topics which are of particular interest to them.
In addition to discussing traditional serial algorithms, we will discuss parallel computation. Students will be given accounts on Utah's Center for High Performance Computing facilities.
It is assumed (indeed, required) that the students have some familiarity with Unix, C, C++ or Fortran programming languages, and Maple. No familiarity with parallel systems is expected.
Some links related to the course are listed below. Often useful tidbits, perhaps even direct hints which are relevant to homework assignments, will appear at the URL entitled Web Resources. Please monitor the Announcements page for updates regarding the more bureaucratic aspects of the course. <>