Math 310
See syllabus .
Fall 2002 ( Ed Bueler )

Assignment #4

Due Friday 10/18

Exam # 1 is on Monday 10/14.
 It will cover sections 1.1, 1.2, 2.1, 2.2, 2.3, 3.1, 3.2, 3.3.
The problems will be like the easier homework questions from those sections.

Note:   You may always use Matlab to do computations onthe homework.  However, when explanations, arguments, "show that ..." are called for, you need to include your written answer in English.  Also, please try to minimize the amount of paper your solution uses.  You can copy a few numbers by hand, sometimes, to save a page-long printout with only those few numbers.  Suggestion: Print out programs and write in the white space.

From text (Cheney&Kincaid, 4th ed.):

§2.3: Problems: #  1, 2, 5, 29
§2.3: Computer Problems: # 1
§4.1: Problems: #  1, 3, 4, 8, 16, 27    Then figure out how to use Matlab's  polyfit and  polyval to do the following problems:
(§4.1) Problem A:   Plot   f(x)=sin x  and the polynomial interpolant of degree 4 which has the values yj=sin xj for x0=-1, x1=-.5, x2=0, x3=.5, x4=1.  Plot on the same graph.  Make smooth graphs of both  f  and the polynomial, and make the interpolating points clear.  That is, match the plot below.
(§4.1) Problem B:   Plot   f(x)=(1+16 x^2)^(-1)  and the polynomial interpolant of degree 20 which has the values yj=f(xj)  for xj=-1+(0.1)j,  j=0,1,...,20.  Make a plot like for problem A.
(§4.1) Problem C:   Plot   f(x)=(1+16 x^2)^(-1)  and the polynomial interpolant of degree 20 which has the values yj=f(xj)  for xj=cos(j pi/20),  j=0,1,...,20.  Make a plot like for problem A.
(§4.2) Problem D:   Explain the difference between problems B and C in terms of the Interpolation Errors Theorem I on page 164.
(§1.2) Problem E:  Use Taylor's Theorem to find in what interval around  c=0  the function  f(x)=(1+x)^(1/2)  can be estimated to within 10^(-8)  by the linear function  L(x)=1 + (1/2) x.  
(§1.2) Problem F:  Use Taylor's Theorem to estimate the maximum error used in computing  erf(x)  by  the series  (2/sqrt(pi))(x-(1/3)x^3)   for  x  in the interval  [-1,1].  Compare the the actual error computed from Matlab.  Plot both erf(x) and the series on this interval.

Office: Chapman 301C.  Office hours online.   Contact:  474-7693,  ffelb@uaf.eduwww.math.uaf.edu/~bueler/