MATH 3607 FINAL EXAM
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MATH 3607 FINAL EXAM
Due: Friday, December 8th by 5:00 pm. Please upload a scanned copy on Gradescope and make sure all pages are readable.
Problems marked with , are to be done by hand; those marked with 口 are to be solved using a computer.
When asked to write a MATLAB function include it at the end of your live script.
Problem 1. To solve the linear system Ax = b, where A ∈ Rn ×n , b ∈ Rn, the following iterative method is proposed:
xk+1) = −ai,1x k) − · · · − ai,i−1x)1 + (1 − ai,i)xk) − ai,i+1x1 − · · · − ai,nx) + bi
for 1 ≤ i ≤ n and a i,i 0 for all i.
(a) , Write the iteration in matrix form. That is, write the iteration as
x(k+1) = TRx(k) + cR ,
where TR ∈ Rn ×n , cR ∈ Rn. Give TR, cR in terms of A and b.
(b) , Suppose that A is symmetric and positive definite, then it is known that the eigenvalues of A are real and positive. Prove that the spectral radius of the iteration matrix, TR, is
ρ(TR) = max{| 1 − λ1 | , | 1 − λn |},
where λ1 and λn are, respectively, the biggest and smallest eigenvalues of A.
(c) , Show that, when A is symmetric and positive definite, the iteration converges if and only if λ1 < 2.
(d) 口 Write a MATLAB function myiteration which performs the above iteration method in matrix form. Your function should have as inputs a matrix A, a vector b, the initial guess vector x0 , the tolerance TOL, and an integer max__iterations. As a stopping criteria,
use the condition
∥x(n) − x(n− 1)∥∞
∥x(n) ∥∞
2x1 − x2 = 3
−x1 + 2x2 − x3 = 3
− x2 + 2x3 = 1
and
1 2 = 2
2 3 = 1
1 3 = 4
Use your program to approximate the solution the systems with x(0) = 0 and tolerance within 10−5 in the 钝-norm. For which system does the iteration method give a good approximation within 25 iterations? Justify your answer.
Problem 2.
(a) , Let n > 1 be a positive integer. Use Newton’s method to produce a quadratically convergent method for calculating the nth root of a positive number a. Prove quadratic convergence.
(b) 口 Consider the following iterative method to find a root p of a function f. First choose initial approximations p0 and p1 with f(p0 ) · f(p1 ) < 0. The approximation p2 is chosen as in the Secant method, as the x-intercept of the line joining (p0 , f(p0 )) and (p1 , f(p1 )). To decide which secant line to use to compute p3 consider f(p2 ) · f(p1 ):
. If f(p2 ) · f(p1 ) < 0, then there is a root between p2 and p1 , and choose p3 as the x-intercept of the line joining (p1 , f(p1 )) and (p2 , f(p2 )).
. if not, choose p3 as the x-intercept of the line joining (p0 , f(p0 ) and (p2 , f(p2 )), and then interchange the the indices on p0 and p1 .
. In a similar manner, once p3 is found, the sign of f(p3 ) · f(p2 ) determines whether we use p2 and p3 or p3 and p1 to compute p4. If p3 and p1 are used, a relabeling of p2 and p1 is performed.
Write a MATLAB function myrootmethod.m which implements the above algorithm. Use your method to find a root of f(x) = cos (x) — x with p0 = 0.5 and p1 = π/4 and compare the approximations to those obtained using Newton’s Method and the Secant method.
Problem 3. Consider the problem of finding a degree at most 4 polynomial interpolant for the function f(x) = ex on the interval [ — 1, 1].
(a) , Let P4(x) be the degree at most 4 Lagrange interpolating polynomial interpolating f at the nodes x0 = — 1, x1 = — 1/2, x2 = 0, x3 = 1/2, x4 = 1. Find the maximum error when P4(x) is used to approximate f(x) for x e [ — 1, 1].
(b) , Let Q4 (x) be the interpolating polynomial of degree at most 4 with nodes at the zeros
of T(˜)5 (x). Find the maximum error when Q4 (x) is used to approximate f(x) for x e [ — 1, 1].
Comment on how this compares to part(a).
(c) 口 In one figure plot Q4 , f, and the Tchebyshev nodes.
Problem 4.
(a) , Use Newton’s divided differences to find the Hermite cubic polynomial that interpolates the following data
x |
0 |
1 |
f(x) |
0 |
0 |
f’(x) |
1 |
1 |
(b) 口 The following two point sets define the top and bottom of a flying saucer shape. Top:
0 |
0.51 |
0.96 |
1.06 |
1.29 |
1.55 |
1.73 |
2.13 |
2.61 |
|
|
0 |
0.16 |
0.16 |
0.43 |
0.62 |
0.48 |
0.19 |
0.18 |
0 |
Bottom:
x |
0 |
0.58 |
1.04 |
1.25 |
1.56 |
1.76 |
2.19 |
2.61 |
|
0 |
-0.16 |
-0.15 |
-0.30 |
-0.29 |
-0.12 |
-0.12 |
0 |
Use piecewise cubic interpolation to make a picture of the flying saucer.
Problem 5. Consider the set of m points {(xi, yi) | i = 1, . . . , m} which are arranged in a nearly circular pattern. Suppose these coordinates are saved in MATLAB as column vectors x and y and we would like to determine the center (c1 , c2 ) and the radius r of the circle that best fits the data.
(a) , Transform the nonlinear problem into a linear least squares problem by rearranging the equation
(x − c1 )2 + (y − c2 )2 = r2
and introducing a new undetermined coefficient. Express the linear least squares problem as Vc‘‘ = ”b.
(b) , 口 Write a MATLAB program which calculates c1 , c2 , and r based to the linear least squares formulation derived in part (a). Write your program in four lines of MATLAB code: one line to construct V , one line to construct b, one line to solve for c, and the last line to determine r.
(c) 口 Download circ.mat and load the the data in MATLAB using load('circle.mat'). Run your program written in (b) using the loaded data. Plot the fitting circle and the data points together.
2023-11-24