This preview shows page 1. Sign up to view the full content.
Unformatted text preview: 08/11/2000 FRI 11:33 FAX 6434330 MOFFITT LIBRARY M. Rieffel Mathematics 128A EnaJBcamjmﬂon December 18, 1997 SHOW YOUR WORK COMYANDNEATLY. Total points = 140. 1.Let p bethepolynomialwhichinterpolates f(x)=1/; atlhe
15 points 4, 9, and 16. ﬁrmlanupper boundfor |f(x)-p(x)l
for 9 s x s 10. Justifyyour answer. 3 2'
2.121: A=[_l 4]. 15 a) Find the LU decomposition of A. . 9 b) Use the LU decomposition of A to solve the equation Ax = b 1
for b=[_1] . 1 5 c) Obtain an estimate for the condition number of A for a norm
of your choice (speciﬁrit). Justify your answer. 3. a) Describe brieﬂy the strategy for deriving the Runge—Kutta '3 7
4 methods for solving ODE's. b) Explain precisely the derivation, following your strategy above,
14 of the second order RungeKutta meﬂmd c) Deﬁne what is meant the local truncation error for a
3 single-step method for solving ODE‘S. ' d) Show that the local truncation error for the above method is
12 of order n3 . . (over) 001 08/11/2000 FRI 11:33 FAX 6434330 MUFFITT LIBRARY 4. a) Give a brief but precise geometric explanation of how one
8 obtains the formula for the Newton-Raphson method for ﬁnding the zeros of a function. (Include a precise statement of
the formula). h) Deﬁne precisely what it means for arconvergent sequence of
3 numbers to converge quadratically. c) Show precisely why the Newton-Raphson method often gives
1 2 quadratic convergence. - S. The Laguerre polynomials, L. , are orthogonal polynomials on
15 [0, +co) for the weight function w(1)= e" . They satisfy the
beautiful recursion relation A
(n+11L,+1(x) = (211-4-1— x).L.(x) - alrﬁx),
while Lem-=1 and le-l-x.
Derive the Gaussian two-point integration formula for
g f(x)e"‘dx . 6. Suppose you are equipped with a pocket calculator with trig
1 5 functions (but without an integrate key), and you need to ﬁnd
ﬁmxzmx withan errorof < 10". Brphinpredsdyhow
you Would proceed to do this, eg. at what points you would
evaluate 009(3) and what you would do with the values. (You do not need to carry out the computation.) Prove that your
procedure will give an answer of the required accuracy. 002 ...
View Full Document
This note was uploaded on 05/17/2009 for the course MATH 128A taught by Professor Rieffel during the Spring '08 term at University of California, Berkeley.
- Spring '08