Error Analysis and Convergence of Iterative Method
Error Analysis and Convergence of Iterative Method
Muhammad Irfan
School of Electrical Engineering and Computer Science (SEECS)
National University of Sciences and Technology,
Islamabad, Pakistan
Lecture (1 and 2)
Outlines of Lecture
• Mathematical Preliminaries
Main Objectives
Numerical Method
Definition
It is the branch of mathematics which deals with the solution of
mathematical problem, having initial information/ conditions in
numerical form and results are required in numerical form. The
technique applied to achieve the results are called numerical
methods.
This course has some distinct features. It can give the numerical
solution of the problems.
•Which have no analytical method for their solution.
•The analytical methods exists but are very cumbersome to use.
•The results up to desired accuracy are obtained rapidly.
Error Analysis
Definition
Error is a term used to denote the amount by which an
approximation fails to equal the exact solution.
Sources of Error
Truncation Errors:
Often an approximation is used in place of an exact mathematical
procedure. For instance consider the taylor series expansion of sin x
is
3 5 7
sin x = x − x3! + x5! − x7! + .....
Practically we cannot use all of the infinite number of terms in the
series for computing the sin x. we usually terminate the process
after the certain number of terms. The error that results due to
such a termination or truncation is called truncation error.
Example:
Expand f (x) = e x
x2 x3 x4 x5
f (x) = e x = 1 + x + + + + .....
2! 3! 4! 5!
Muhammad Irfan Introduction to Numerical Method and Preliminaries
Introduction to Numerical Method and Preliminaries 8/23
Preliminaries
Relative Error
Definition
If a is an approximate value of a quantity whose exact value is a
then relative error r of a is defined by
|a−a| ||
| r |= =
| true value | |a|
Example:√
Consider
√ 2 = 1.414213.... upto four decimal places then
2 = 1.4142 + errors
| error | =| 1.4142 − 1.41421 | = 0.00001, taking 1.4142 as true
value or exact value
Hence, r = 0.00001
1.4142
Muhammad Irfan Introduction to Numerical Method and Preliminaries
Introduction to Numerical Method and Preliminaries 12/23
Preliminaries
Absolute Error
Definition
If a is an approximate value of a quantity whose exact value is ’a’
then the difference =a − a is called absolute error of ’a’.
Error Bound
Definition
It is a number β for a such that | a − a |≤ β or | |≤ β.
Example:
√ If β = 10−4 and a = 1.4142 is an approximation to
a = 2 = 1.4142135... then the error is
=| a − a |= 0.0000135 ≤ β.
Error
Definition
Local Error:
The Local error is the error introduced during one operation of the
iterative process. (This is the error after first step.)
1 = x(to ) − x
Definition
Local Error:
Global error measures the difference between the computed
solution at a given time and the true solution of the initial value
problem that we are trying to solve.
n = x(tn ) − x
Muhammad Irfan Introduction to Numerical Method and Preliminaries
Introduction to Numerical Method and Preliminaries 15/23
Preliminaries
Definition
Precision:
Precision mean how close are the measurements obtained from
successive iterations.
or
Precision is how consistent results are when measurements are
repeated. Precise values differ from each other because of random
error, which is a form of observational error.
Definition
Example:
If you take measurements of the mass of a 50.0 gram standard
sample and get values of 47.5, 47.6, 47.5, and 47.7 grams, your
scale is precise, but not very accurate. The average of your
measurements is 47.6, which is lower than the true value. Yet,
your measurements were consistent. If your scale gives you values
of 49.8, 50.5, 51.0,and 49.6, it is more accurate than the first
balance but not as precise. The average of the measurements is
50.2, but there is a much larger range between them. The more
precise scale would be better to use in the lab, providing you made
an adjustment for its error. In other words, it’s better to calibrate
a precise instrument than to use an imprecise, yet accurate one.
Numerical Algorithm
Definition
A complete set of procedures which gives an approximate solution
to a mathematical problem.
Definition
Criteria for a good method
1) Number of computations i.e. Addition, Subtraction,
Multiplication and Division.
2) Applicable to a class of problems.
3) Speed of convergence.
4) Error management.
5) Stability.
Definition
Stable Algorithm:
Algorithm for which the cumulative effect of errors is limited, so
that a useful result is generated is called stable algorithm.
Otherwise Unstable.
Definition
Numerical Stability:
Numerical stability is about how a numerical scheme propogate
error.
Definition
Algorithm of Iteration Method:
A specific way of implementation of an iteration method, including
to termination criteria is called algorithm of an iteration method.
In the problem of finding the solution of an equation, an iteration
method uses as initial guess to generate successive approximation
to the solution.
Definition
Why we use numerical iterative methods for solving
equations?
As analytic solutions are often either too tiresome or simply do not
exist, we need to find an approximate method of solution. This is
where numerical analysis comes into picture.
Definition
Rate of Convergence of an Iterative Method:
Suppose that the sequence (xk ) converges to ’r’ then the sequence
is said to converge to ’r’ with order of convergence ’a’ if there exist
a positive constant ’p’ such that
limk→∞ |x|xk+1 −r |
k −r |
k+1
α = limk→∞ α = p
k
Thus if α = 1, the convergence is linear. If α = 2, the convergence
is quadratic and so on.Where the number a is called convergence
factor.
Muhammad Irfan Introduction to Numerical Method and Preliminaries
Introduction to Numerical Method and Preliminaries 22/23
Preliminaries
Example:
2x + y = 4, 2x + 1.01y = 4.02 has the exact solution
x = 1, y = 2
If we do small change in the equation 2x + y = 3.8,
2.02x + y = 4.02
It is easily verified that the exact solution of this system is
x = 11, y = −18.2.
This solution differs greatly from the solution of the first system.
Both these systems are said to be ill-conditioned.
Muhammad Irfan Introduction to Numerical Method and Preliminaries
Introduction to Numerical Method and Preliminaries 23/23
Preliminaries