G. R. Ierley
This course for first-year graduate students is intended to supplement the analytic skills mastered in undergraduate mathematics courses by developing each student's intuition for the ramifications of the transition from analysis to numerical solution. Numerical solution nearly always involves approximations above and beyond those already implicit in any physical model hence it is important to understand topics such as discrete mathematics, perturbation theory and asymptotic analysis as they bear on numerical accuracy. Numerical efficiency is often at issue for practical methods of solution hence a second major emphasis in this course is on the importance of appropriate analysis for efficient algorithm development. (The relation between algorithmic design and machine architectures, while equally important for applications in large scale computing, is not covered in this course.) This course assumes a reasonable competence with any scientific programming language and a standard background in undergraduate mathematics up to and including ordinary differential equations. Some knowledge of linear algebra is helpful.
4 units, 3 lecture hours/week, Winter Quarter
The text is Sorry, Wrong Number, available from the instructor in PostScript from the TeX manuscript currently under development.