25.08.2021 Views

082-Engineering-Mathematics-Anthony-Croft-Robert-Davison-Martin-Hargreaves-James-Flint-Edisi-5-2017

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

918 Chapter 28 Probability

Engineeringapplication28.10

Redundancyofabinarydatastream

Consider the source of binary digits examined in Engineering application 28.6 and

28.9. The maximum entropy for a binary stream is 1 bit per binary digit. Calculate

the redundancy ineach case.

Solution

For Engineering application 28.6

redundancy = 1 −1 = 0

1

For Engineering application 28.9

redundancy = 1 −0.881

1

= 0.119

Engineeringapplication28.11

Redundancyofacharacterdatastream

A stream of data consists of four characters A, B, C, D with probabilities 0.1, 0.3,

0.2,0.4,respectively. Calculate the redundancy.

Solution

It can be shown that the maximum entropy, H max

, corresponds to the situation in

whichthe probabilityofeachsymbolisthe same, thatis0.25.

H max

= 4 × (−0.25log 2

(0.25)) = 2 bits

Theactual entropy,H act

, isgiven by

H act

= −(0.1log 2

(0.1) +0.3log 2

(0.3) +0.2log 2

(0.2) +0.4log 2

(0.4))

= 1.846bits

redundancy = 2 −1.846

2

= 0.0770

Intheexampleswehaveexaminedsofarwehaveusedthebitastheunitofinformation

because the most common form of digital signalling uses binary digits. When there are

only two possible events it is possible to represent an event by a single binary digit.

However, if there is a larger number of possible events then several binary digits are

neededtorepresentasingleevent.Whencalculatingvaluesforinformationandentropy

in these examples an assumption was made that each event was represented by binary

sequences or codes of the same length. It is only possible to do this efficiently if the

numberofeventsisapowerof2,thatis2,4,8,16, ....Inpracticethisproblemdoesnot

arise because it is more common to produce codes that have a small number of binary

digitsforlikelyeventsandalongnumberofbinarydigitsforunlikelyevents.Thisallows

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!