Sei sulla pagina 1di 17

Numerical Analysis

Course code:GS-301 Instructor Name: M. Shahid Iqbal


Smester: 6
th
Contact info: m.shahid@uog.edu.pk
Office: EE-207 Office Hours: Thursday 12:00:3:00pm

Course Objectives
To enable the students appreciate the
significance of numerical methods for solving
engineering problems.
Text Book
Advance Engineering Mathematics.
Erwin Kreyszig 8
th
Edition
Numerical Analysis
Timothy Sauer
Attendance
A minimum of 70% attendance is required for a
student to be eligible to take the final
examination.
The students with less than 70% of the attendance
in a course shall be given the grade SA (Short
Attendance) in such a course and shall not be
allowed to take its End Term Exams and will
have to reappear in the course to get the required
attendance to be eligible to sit in the exam when
the course is offered the next time.

Grading
The course will be evaluated on the basis of the
following percentage:
Mid Term 25%
Sessional work 25%
Presentation 10%
Assignment 10%
Quizzes 05%
Final term 50%

First week Plan
Introduction
Floating point
Round off
Error Propagation
Floating point form of numbers
A floating point number consists of three parts
Sign (+ or -)
Mantissa (contains a string of digits)
Exponent

Levels of Precision
There are three levels of precision which are
commonly used

Single precision (32 Bits)
Double precision (64 Bits)
Extended precision (80 Bits)
Bits division
Sign Exponent Mantissa
Single precision 1 8 23
Double precision 1 11 52
Extended precision 1 15 64
Bits storage
Double precision 1 will be stored like
+1.0000..0 X 2^0
The next floating point is
+1.0000..001 X 2^0
Or 1+2^-52
Machine epsilon denoted by is the
distance between 1 and the smallest
floating point number greater than 1.
=2^-52

Decimal notation
Two notations are used
Fixed point notation
Floating point notation
Fixed point System
In fixed point system all digits are given with a
fixed number of places i.e.
621.358
0.013
1.000
Impractical in most scientific calculations.
Floating point
In floating point system number of significant
digits is kept fixed and the decimal point is
floating i.e.

0.6238 X 10^3
0.1714 X 10^-13
Or equally 0.6238E03, 0.1714E-13
Significant digits
Significant digit of a number C is any given
digit of C except possibly for zero to the left
of the first non zero digit. i.e.
1360, 1.360 has 4 significant digits.
Under Flow and Over flow
IEEE floating point standard for single precision
is about -38 < e < 38.
IEEE floating point standard for double precision
is about -308 < e < 308.
If in a computation a number exists outside the
smaller limit it is called under flow.
If it exists outside the upper limit it is called
overflow.

Round Off
Chopping
Just remove the extra digits with out any change
in the remaining ones.
Chopping cause more error.
Round off
Discard (k+1)th and subsequent digits.
If the discarded digit is greater than the half
unit add 1 to the kth digit.
If the discarded digit is smaller than the half
unit the kth digit will remain unchanged.
If it is equal to half unit round off to the nearest
even digit. i.e. 3.45 and 3.55 will be 3.4 and
3.6.
Examples
Write 23.49,-302.867,0.000527532 and -25700
in floating point form rounded to 4S.
0.81534/(35.724-35.596) this is a 5S value
solve it. After that round it to 4S and solve
again now round off to 3S to the values of 4S
and do the same.
Repeat it for 2S and observe the round off
error.

Potrebbero piacerti anche