Sie sind auf Seite 1von 4

18.

100A PS17
Author: Eric Emer April 29, 2013

Reading 17 and 18 of Mattuck Collaborators: None.

1
1.1

Exercise 17.2.2
a
P (x) = a0 + a1 x + a2 x2 + . . . + an xn

apply x = u + a: P (u + a) = a0 + a1 (u + a) + a2 (u + a)2 + . . . + an (u + a)n apply u = x a P (x) = c0 + c1 (x a) + c2 (x a)2 + . . . + cn (x a)n Proof. We wish to show that the function above P (x) at point a is the same as the Taylor polynomial. Suppose we want to nd the Taylor Polynomial for a function f (x) = P (x) at the point a. f (x) = c0 + c1 (x a) + c2 (x a)2 + . . . + cn (x a)n f (a) = c0 We examine the rst derivative f (x) = c1 + 2c2 (x a) + 4c4 (x a)2 + . . . + ncn1 (x a)n1 f (a) = c1 We examine the second derivative: f (x) = 2c2 + (3)(2)c3 (x a) + . . . + n(n 1)cn (x a)n2 1 c2 = f (a) 2 yielding the general formula: ck = 1 (k) f (a) k!

Tn (x) =
k=0

f (k) (a) (x a)k k!

And we have the Taylor polynomial, which is indeed the same as the resulting polynomial.

1.2
Blah.

Exercise 17.3.2

We examine that this is really just the Taylor Polynomial for f (x) = cos x, with n = 2 evaluated at a = 0. This gives us remainder function: R2 (x) = sin x f (3) (x) (x a)3 = (x a)3 3! 6

1 x We note that | sin 6 | 6 . We are interested in x, a such that:

sin x (x a)3 | .0001 6 So we apply our estimates, and we assume that a = 0: | 1 | (x a)3 | .0001 6 |(x a)3 | .0006 |(x)3 | .0006 3 |x| + .0006 |x| .0843433 We conclude that this approximation, cos x 1
x2 2 ,

will be valid for x [.0843433, .0843433].

3
3.1

Problem 17-1
a

Proof. We wish to show that a polynomial P (x) has a k -fold zero at a P (a) = P (a) = . . . = P (k1) (a), P (k) (a) = 0. First we show the forward relationship. If a polynomial P (x) has a k -fold zero at a, then: P (x) = (x a)k Q(x), Q(a) = 0 We examine the nth derivative of P (x). P (x) = k (x a)k1 Q(x) + (x a)k Q (x) P (x) = k (k 1)(x a)k2 Q(x) + 2k (x a)k1 Q (x) + (x a)k Q (x) Lets dene the rst term of the derivative as F (x). Where F1 (x) = k (x a)k1 Q(x) and F2 (x) = k (k 1)(x a)k2 Q(x). We can see that: Fn (x) = k (k 1) . . . (k (n 1))(x a)kn Q(x) And so, Fk (x) = k !(x a)0 Q(x) = k !Q(x) Going back to P (x), P (x), we see that in fact all of the terms when evaluated for P (a) are zero. This is because they all contain a power of (x a). The only time when a non-positive power of (x a) occurs is at the k -th derivative, when (x a)0 , is a power, and we get a single non-zero term in P k (a). Such that, P (k) (a) = k !Q(a), and we assume that Q(a) = 0 from the denition of k -fold zero.

We also examine the backwards relationship. Suppose a polynomial function P (x) has the property: P (a) = P (a) = . . . = P (k1) (a), P (k) (a) = 0 Suppose for the sake of contradiction, that P (x) does not have a k -fold zero. Therefore, we can write P (x) = (x a)Q(x), where Q(a) = 0. However, as show earlier, the terms in P (k) (a) all contain powers of (x a) and are therefore zero, except for Fk (x) = k !Q(x). So we get P (k) (a) = k !Q(a). Since Q(a) = 0, we get P k (a) = 0, which is a contradiction.

3.2

Using the above theorem, we apply it to f (x) = 2x3 bx2 + 1. At 2nd fold zero, c: f (c) = f (c) = 0 f (c) = 2c3 bc2 + 1 = 0 f (c) = 6c2 2bc = 0 Solving the linear system, we get: b = 3 and c = 1. The only solution. 3

3.3

The connection is clear. If there is an interval (a, b) such that f (k) (a) = 0 and f (k) (b) = 0, then there must be some c (a, b) such that f (k+1) (c) = 0.

Das könnte Ihnen auch gefallen