Numerical Analysis I


Course Syllabus   Homework  
Maple Worksheets and Projects   Dr. Rayevskaya's Schedule



Announcements:      

Maple intro: http://math-cs.cns.uni.edu/~rayevska/calc1_project.mws

Reminder: no class 9-05-05 - Labor Day
Test 9-30, open book!
Review homework problems (see the list above)

Important points include, but not limited to:
1.1 Taylor polynomial, error term, applications
1.2 Round-off error, mantissa, chopping, rounding, absolute and relative error
1.3 Linear, exponential algorithm, rate of convergence
2.1 Bisection algorithm, error
2.2 Fixed-point iteration, error bounds (various)
2.3 Newton's algorithm, Secant method
2.4 Linear and quadratic methods
also: zero of multiplicity k, stopping techniques (familiarity with).

Program Disc

test 1
test 2


Project 1


Take-home portion of the final:
1. Interpolation of f(x) on [a,b].
(a) Lagrange interpolate f on [a,b] using n nodes.
(b) Hermite interpolate f on [a,b] using n nodes (a degree 2n-1 algebraic polynimial).
2. Consider the IVP y'=f(t,y) on [a,b] with y(a)=alfa. Solve numerically, with n steps (or a given h), using the following methods:
(a) Euler
(b) (Runge-Kutta) Midpoint
(c) Runge-Kutta 4th order
(d) Adams-Bashford two step
note: 2(d) The method is given in the book, so you can just program it from there. But you also need to derive/justify it, i.e. compute the coefficients you use for Wi and Wi-1 in the formula.
Update: demonstrate solutions to 1(a-b) using a concrete function, for example f(x)=e^x, on [-1,1]. In class I suggested you use five nodes for Lagrange interpolation and degree 3 (thus, two nodes, with function and derivative values) for Hermite interpolation.
Also, demonstrate solutions to problem 2 using the system y'=1+y/t, t in [1,2], n=10 (or h=0.1).



Email Dr. Rayevskaya : rayevska@math.uni.edu