Day |
Section |
Description |
1. |
1.1 |
Numbers, evaluation of expressions, and automatic differentiation. |
2. |
1.1 |
Explanation of software for automatic differentiation (excerpts from
Rigorous
Global Search: Continuous Problems, Chapter 2.) |
3. |
1.2 |
Floating point numbers, roundoff error. |
4. |
1.3 |
numerical stability. |
5. |
1.4 |
Error propagation and condition numbers. |
6. |
1.5 |
Interval arithmetic. |
7. |
1.5 |
Interval evaluation of expressions; software tools (excerpts from Rigorous
Global Search: Continuous Problems, Chapter 2.) |
8. |
2.1 |
Gaussian elimination. |
9. |
2.1 |
More on Gaussian elimination. |
10. |
2.2 |
Structured systems of equations |
11. |
2.3 |
Rounding error analysis for Gaussian elimination; pivoting |
12. |
2.4 |
Vector and matrix norms |
13. |
2.4 |
More on vector and matrix norms |
14. |
2.5 |
Condition numbers |
15. |
2.7 |
Error bounds for linear systems |
16. |
3.1 |
Polynomial interpolation |
17. |
3.1 |
More on polynomial interpolation. |
18. |
3.2 |
Numerical differentiation. |
19 |
3.3 |
Cubic splines. |
20. |
3.4 |
Approximation by splines |
21. |
3.5 |
Radial basis functions |
22. |
5.1 |
The secant method; linear, quadratic, and superlinear convergence |
23. |
5.2-5.3 |
Bisection methods and bisection methods for eigenvalue problems |
24. |
5.4 |
Convergence order |
25. |
5.5 |
Error analysis; interval Newton method |
26. |
5.7 |
Newton's method |
27. |
|
Line searches: supplement |
28. |
6.1 |
Systems of nonlinear equations: preliminaries. |
29. |
6.2 |
Theory of Newton's method. |
30. |
6.3 |
Error analysis for the multivariate Newton's method; interval Newton
method; supplement with section 1.5 of Rigorous
Global Search: Continuous Problems. |
31. |
6.4 |
Other methods |
32. |
4.1 |
Quadrature formula theory |
33. |
4.2 |
Gaussian quadrature |
35. |
4.3 |
The trapezoidal rule and extrapolation. Supplement: The trapezoidal
rule for periodic functions. |
36. |
4.4 |
Adaptive integration. |
37. |
Ch. 6 |
Time permitting, additional topics will be covered. The plan
for next semester is to cover optimization, ordinary and partial differential
equations. (There are 41 meeting periods in the fall semester.) |