You are on page 1of 3

6CCM359A: Numerical Differentiation

Introduction

1. Use the forward and backward-difference formulas to approximate the derivatives in the
following tables. Knowing the functions from which the tables were generated, calculate
the actual errors and find error bounds using the error formulas.

(I) Table:

x f (x) f 0 (x)
0.5 0.4794
0.6 0.5646
0.7 0.6642

Function: f (x) = sin(x)

(II) Table:

x f (x) f 0 (x)
0.0 0.00000
0.2 0.74140
0.4 1.3718

Function: f (x) = ex − 2x2 + 3x − 1

(III) Table:

x f (x) f 0 (x)
-0.3 1.9507
-0.2 2.0421
-0.1 2.0601

Function: f (x) = 2 cos(2x) − x

(IV) Table:

x f (x) f 0 (x)
1.0 1.0000
1.2 1.2625
1.4 1.6595

Function: f (x) = x2 ln(x) + 1

2. Use the most appropriate three-point formula to approximate the derivatives in the fol-
lowing tables. Knowing the functions from which the tables were generated, calculate
the actual errors and find error bounds using the error formulas.

1
(I) Table:

x f (x) f 0 (x)
1.1 9.025013
1.2 11.02318
1.3 13.46374
1.4 16.44465

Function: f (x) = e2x

(II) Table:

x f (x) f 0 (x)
-0.3 -0.27652
-0.2 -0.25074
-0.1 -0.16134
0 0

Function: f (x) = e2x − cos(2x)

(III) Table:

x f (x) f 0 (x)
2.0 3.6887983
2.1 3.6905701
2.2 3.6688192
2.3 3.6245909

Function: f (x) = 2(ln(x))2 + 3 sin(x)

Richardson’s extrapolation

3. Use Richardson’s extrapolation techniques as explain during lectures to determine N3 (h)


as an approximation for f 0 (x0 ) for the following functions and step sizes. Calculate the
exact value for the derivative and the actual error.

(I) f (x) = ln(x), x0 = 1, h = 0.4

(II) f (x) = x + ex , x0 = 0.0, h = 0.4

(III) f (x) = 2x sin(x), x0 = 1.05, h = 0.4

(IV) f (x) = x3 cos(x), x0 = 2.3, h = 0.4

2
4. Suppose that N (h) is an approximation to M for h > 0 and that the error is given by

M = N (h) + K1 h + K2 h2 + K3 h3 + · · ·

for some constants Ki i = 1, 2, . . .. Use the values N (h), N (h/3) and N (h/9) to produce
an O(h3 ) approximation for M .
5. Suppose that N (h) is an approximation to M for h > 0 and that the error is given by

M = N (h) + K1 h2 + K2 h4 + K3 h6 + · · ·

for some constants Ki i = 1, 2, . . .. Use the values N (h), N (h/3) and N (h/9) to produce
an O(h6 ) approximation for M .
IPC—30 September 2010

You might also like