Scribd
Upload
20 views

Numerical Methods I - Error Propagation

Uploaded by

letif12aa
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
20 views

Numerical Methods I - Error Propagation

Uploaded by

letif12aa
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 30

Azərbaycan Dövlət

Neft və Sənaye
Universiteti
Numerical
Methods I
Error Propagation
The purpose of this presentation is to
study how errors in numbers can
propagate through mathematical
functions.
For example, if we multiply two
numbers that have errors, we would
like to estimate the error in the
product.
Suppose that we have a function that
is dependent on a single independent
Assume that is an approximation of .
We, therefore, would like to assess the
effect of the discrepancy between and
on the value of the function; i.e. we
would like to estimate

The problem with evaluating is that


is unknown because is unknown.

We can overcome this difficulty if is


close to and is continuous and
If these conditions hold, a Taylor
series can be employed to compute
near , as in

Dropping the second- and higher-order


terms and rearranging yields

or
Here

represents an estimate of the error of


the function ;

and

represents an estimate of the error of .


Eq. provides the capability to
approximate the error in given the
derivative of the function and an
estimate of the error in the
independent
variable.
: Graphical
depiction of
first-order error
propagation.
1. In this example, we will evaluate
error propagation in a function of a
single variable.

Given a value of with an error of ,


estimate the resulting error in the
function .
.
Using Eq. , we obtain
.
Because , we predict

or that the true value lies between


and .
In fact, if were actually , the function
could be evaluated as ; and if were , it
would be .
For this case, the first-order error
analysis provides a fairly close
The foregoing approach can be
generalized to functions that are
dependent on more than one
independent variable.
This is accomplished with a
multivariable version of the Taylor
series.
For example, let’s say that we have a
function of two independent variables
and .
The Taylor series then can be written
as

Here all partial derivatives are


evaluated at the base point .
If all second-order and higher terms
are dropped, Eq. can be solved for

where
and
are estimates of the errors in and ,
respectively.
For independent variables , ,…,
having errors , ,…, , the following
general relationship holds:

Equation can be employed to define


error propagation relationships for
common mathematical operations.
The results are summarized in below.
Operation Estimated Error

Addition

Subtraction

Multiplicati
on

Division

presents estimated error bounds


associated with common mathematical
operations using inexact numbers and

The of a mathematical problem
relates to its sensitivity to changes in
its input values.

We say that a computation is if the


uncertainty of the input values is
grossly magnified by the numerical
method.
These ideas can be studied using a
first-order Taylor series: .
This relationship can be employed to
estimate the relative error of as in

The relative error of is given by


A can be defined as the ratio of these
relative errors

The condition number provides a


measure of the extent to which an
uncertainty in is magnified by .

A value of tells us that the function’s


relative error is identical to the
A value greater than tells us that the
relative error is amplified, whereas a
value less than tells us that it is
attenuated.
Functions with very large condition
numbers are said to be .
Any combination of factors in Eq. that
increases the numerical value of the
condition number will tend to magnify
uncertainties in the computation of .
.
In this example, we compute and
interpret the condition number for

and
.
The condition number is computed as

For , we obtain:

Thus, the function is ill-conditioned.


.
For , the situation is even worse:

For this case, the major cause of ill-


conditioning appears to be the
derivative.
This makes sense because in the
vicinity of , the tangent approaches
both positive and negative infinity. ◄
.
In this example, we calculate error
propagation in a multivariable
function.
The deflection of the top of a sailboat
mast is

where is a uniform side loading , is


the height , is the modulus of
elasticity , and is the moment of
.
Estimate the error in given the
following data:
.
Employing Eq. gives

or

Substituting the appropriate values


gives
.
Therefore

In other words, is between and .

The validity of these estimates can be


verified by substituting the extreme
values for the variables into the
equation to generate an exact
.
These values are

Thus, the first-order estimates are


reasonably close to the exact values.
Thank you very much for
attention!
text

text

text
(continued)
text

text

text
(continued)
text

text

text

You might also like