0% found this document useful (0 votes)
69 views38 pages

CL 244 - Module - 1

The document is a lecture on numerical analysis given by Sarika Mehra. It discusses sources of approximation in numerical modeling from empirical measurements and discretization. It also covers error types like relative and absolute error. Different types of errors from computation and rounding are examined. Properties of floating point representation and machine precision are outlined.

Uploaded by

Anuj Bhawsar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
69 views38 pages

CL 244 - Module - 1

The document is a lecture on numerical analysis given by Sarika Mehra. It discusses sources of approximation in numerical modeling from empirical measurements and discretization. It also covers error types like relative and absolute error. Different types of errors from computation and rounding are examined. Properties of floating point representation and machine precision are outlined.

Uploaded by

Anuj Bhawsar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 38

CL244: Introduction to Numerical

Analysis

Sarika Mehra

Dept. of Chemical Engineering

I.I.T. Bombay, Mumbai 400076

Email: [email protected]

Sources of Approximations

• Modeling.

• Empirical measurements or input data

• Previous computations that are input.

• Truncation or discretization.

• Rounding. Numerical

error

CL244: Introduction to numerical analysis 2


Sources of Approximations

CL244: Introduction to numerical analysis 3


Significant Figures

Number of certain digits + 1

estimated.

What is the speed of

the car ?

Chapra et al.

Number of significant figures in each of these numbers?

0.0000145, 0.000145

CL244: Introduction to numerical analysis 4


Error Types

• Precision vs. Accuracy.

• Absolute vs. Relative Error.

• Data Error and Computation Error.

• Forward vs. Backward Error.

• Truncation vs Round-off Error.

CL244: Introduction to numerical analysis 5


Error Types

• Absolute vs. Relative Error.

CL244: Introduction to numerical analysis 6


Error Types

• Precision vs. Accuracy.

• Absolute vs. Relative Error.

• Data Error and Computation Error.

• Forward vs. Backward Error.

• Truncation vs Round-off Error.

CL244: Introduction to numerical analysis 5


Computational Error

• Computational error is the sum of truncation

error and rounding error.

• Truncation error is due to difference in

algorithm, arithmetic is exact.

• Round-error is due to limited precision

arithmetic.

8
CL244: Introduction to numerical analysis

Error Types

• Truncation vs Round-off Error.

CL244: Introduction to numerical analysis 7


Floating-Point Representation

CL244: Introduction to numerical analysis 9

Floating-Point Representation

CL244: Introduction to numerical analysis 9

Properties of Floating-point Representation

Scientific computing by Heath

• Not all numbers have a representation.

• Floating point numbers are not uniformly distributed.

• Overflow and Underflow errors.

• Interval between numbers increases as the numbers grow in

magnitude.

12
CL244: Introduction to numerical analysis

CL24: Introduction to numerical

11
analysis

Round-off Errors

Chapra and Canale

CL244: Introduction to numerical analysis 13


Machine Precision/Epsilon/Unit

Round-off

Alternate definitions:

Distance from 1 to the next possible number.

Smallest number such that Fl(1+e)>1.

Try

1-3*(4/3-1)

14
CL244: Introduction to numerical analysis

Machine Precision/Epsilon/Unit

Round-off

1-3*(4/3-1) = 2 * 10-16

15
CL244: Introduction to numerical analysis

CL24: Introduction to numerical

11
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

Floating-Point Representation

b=2

Single-precision (32 bits)

P=24, L=-126, U=127.

Double-precision (64 bits)

P=53, L=-1022, U=1023.

CL244: Introduction to numerical analysis 10

Floating-Point Representation

b=2

Single-precision (32 bits)

P=24, L=-126, U=127.

Double-precision (64 bits)

P=53, L=-1022, U=1023.

CL244: Introduction to numerical analysis 11

Not stored

1038

11
CL244: Introduction to numerical analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

CL24: Introduction to numerical

16
analysis

You might also like