zProbabilityDistributions

423

= n(n - 1) p+ np
Theformula (17) of Section 7.2.1 now gives the variance of the
binomial distribution as

= Elk- '= n(n - 1) p + np - n'p?

= np(l - p) = npq. (4)

Therefore, the standard deviation of the binomial

distribution is

G = Vnpq ... (5)

Bxample 1. Let x be a binomially distributed random variable

with mean 2 and standard deviation 2/N3. Find the correspond.
ing probability density function.
»Here, H= 2, o= 2/V3. Since =np and o= Vnpg for
the binomial distribution, we have np = 2, npq = 4/3 which
give
4 1 4 1 2 1 2
p= l- q= 3 n= = 6.
q3 np 3 p

Therefore, the probability density function for the given

distribution is b(n, p, x) with n= 6 andp= 1/3, namely
6- x
1
b(6, ) = c,
Example 2. Let x be a binomially distributed random variable
based on 6 repetitions of an experiment. Ifp = 0.3, evaluate the
probabilities P(x = 3), P(xs 3), P(x= 4) and P(x> 4).
»Here, the probability density function is the binomial
probability function
b(6, 0.3, x)= C.(0.3) (1 - 0.3,° -x = b(x), say.
We find that
6- 0
3
b(0) = = 0.1177,
10
424 TU. Engineering Mathematien
6
3 3
b(1) = °c. 10 10 = 0.3025,
6 2

b(2) = 10 =0.3241,
6- 3
3
b(3) =
°C, 10 1
10 =0.1852,
6- 4
3 3
1 = 0.0595.
b(4) = 10 10

Therefore,
P(x = 3) = b(3) = 0.1852,
P(rs 3) = b(0) + b(1) + b(2) + b(3)
= 0.1177+ 0.3025 + 0.3241 + 0.1852 = 0.929:

P(x= 4) = b(4) = 0.0595,

P(x> 4) = 1- P(X< 4)= 1-b(0) + b(1)+ b(2) + b(3) +ha
= 1-(0.1177 +0.3025 +0.3241 +0.1852 +0.0595)

= 1- 0.989 = 0.011.

of
Example 3. When a coin is tossed 4 times, find the probability
and (iii) at least
getting (i)exactly one head, (ii) at most 3 heads,
two heads.
of getting a head in
» Since there are 4 trials and the probability1/2. Therefore, the
each trial is 1/2, we have n= 4 and p =
given by the binomial
probability of getting x heads in 4 trials is
probability function
4
4- X

4,

1
16C)= b(x), say.
Accordingly,
(i) the probability of getting exactly one head is
7.Probability Distributions 425

1 1
b(1) = 16 ('c) = 4
probability of getting atmost three heads is
(ii) the
b (x s 3) = b(0) + b(1) +t b (2) + b (3)
1
*C,+ +

1 15
+
(1+ 4-6+ 4) =
16 16

(i1) the probability of getting atleast two heads is

b (x > 2) = 1-b(x < 2) = 1-b(0)+ b(1)
11
= 1 16
16 16

Example 4. Apair of dice is thrown twice. Find the probability

of scoring 7points (i) once, (ii) twice, and (iii) at least once.
>When a pair of dice is thrown, there are 36 possible outcomes:
(a, b), a =1,2, . 6, b= 1, 2, 6. Here, a is the number shown
on the first die and b is the number on the second. Of these,
a+ b= 7occurs in 6 outcomes, namely (1, 6), (2, 5), (3, 4), (4, 3),
(5,2), (6, 1). Therefore, the probability of scoring 7 points in a
throw is
6 1
p = 36 6

S Such, the probability of getting 7 points x number of times in

twothrows is
b(2, p, x) = 2C, p (1 - p}'
2

a) = bx), say

Accordingly,
(i) the probability of scoring 7 points once is
426 T.U.
Engineering |
b(1) = 5
18 Mathematiety
(ii)the probability of scoring 7points twice is
1
b(2) = c, 36

(ii) the probability of scoring7 pointsat least once ie

5 1 11
b(1) + b(2) = 18 + 36 36

Example 5. The probability that a man aged 60 will

liue
70 is 0.65. What is the probability that out of 10 men, nou toDo94be
60, at least 7 willlive to be 70.
» The probability that a man aged 60 will live to be 70 is given s.
p= 0.65, so that q= 1-p= 0.35.
Therefore, the probability that out of 10 men now aged 60. !
men will live to be 70 is given by the probability function
b( 10, 0.65, x ) = "C (0.65) (0.35)*= b(), say.
Accordingly, the probability that at least 7 of the 10men will
live tobe 70 is
b(x > 7) = b(7) + b(8) + b(9) + b(10)
c,0.65)0.35)°+ c,0.65)° (0.35,
+ "C, 0.65)' 0.35) +(0.65"

= 0.2522 + 0.1757 + 0.0725 + 0.0135

= 0.5139.

instant of
Example 6. The number of telephone lines busy at aninstant10
an
time is a binomial variate with probability 0.2. Ifatthat (i) 5 lines
lines are chosen at randon, what is the probability.
are busy? (ii)at most 2 lines are busy? (iiü) all lines are busy:
The chance that a telephone line is busy is given as
zProbability Distributions 427

that
D= 0.2= 1/5 so q= l-p= 0.8 = 4/5.
Therefore,the probability that x lines out of 10 lines are busy
probability function
is given by the
10 - x

b 10, = b(x), say

Accordingly, the probability that (out of 10), exactly 5 lines are

busy is
5 5
10. 9. 8. 7.6
b(5) = 1: 2 3 4. 5
= 0.02642.

Next, the probability that atmost 2 lines are busy is given by

b(x s 2) = b (0)+ b (1) + b (2)
= + 10c.
= 0.1074 +0.2684+ 0.302= 0.6778.
Lastly, the probability that all the 10 lines are busy is

b(10) =
10C. \10
= 1.024 x 10 -7

Example 7. The probability that a pen manufactured by a

company will be defective is 0.1. If 12 such pens are selected, find
the probability that (i) exactly 2 willbe defective, (i) at least 2 will
be defective, (iii) none will be defective.

The probability that a pen is defective is p = 0.1= 1/10

Since 12 pens are selected, the probability that x pens out of
these 12 are defective is given by the binomial probability
function
12-x
1 1
b 12, 1 = b(x), say.
10 10
428
T.U. Engineering Meathematiey
Therefore:

(i) The probability that exactly 2 pens will be

b(2) =
10 defective is
12! 1
=
X 0.3487 =
10! 2! 100
0.2301.
(ii) The probability that at least 2 pens will be defectiva i.
b(x 2) = 1- b(x <2)
= 1-b(0) + b(1)
0 12
1

= 1 +

9
,12
9
- 1-|0 10

= 1- 0.2824 - 1.2x 0.3138 = 0.3410.

(iii) The probability that none willbe defective is
12 12
9
b(0) = 10 10
= 0.2824.

Example 8. A class of 100 contains 10 bright students. Fe

students from the class are picked at random. Find the following
probabilities: (i) none of the picked is abright student, and () al
the picked are bright students.
Since 10 out of 100students are bright, the probability that i
student chosen is bright is p= 10/100 = 1/10.
Since 5 students are picked, the probability that x students out
of these are bright is given by the binomial probability function

1 1
b|5, = br), say.
C 10 10
Therefore:
() The probability that none of the students picked is brignt
1 Probability
Distributions 429

5
1
b(0) = 1 10

= 0.59.
10
probability that all of the students picked are bright is
(ii))The
b(5) = 10

= 0.00001.
10

Example 9. Amanufacturer of solar heaters claims that 60% of

his heaters work satisfactorily for 10 years. Assuming that his
dlaim is legitimate, find () theprobability that four of five heaters
work satisfactorily for 10 years, and (ii) the probability that
atmost two of five heaters do not work satisfuctorily for 10 years.
» According to the manufacturer's claim, the probability that a
heater works satisfactorily (for 10 years) is p= 60% = 0.6.
Therefore, the probability that x out of 5 heaters work
satisfactorily is given by the probability function
b(5, 0.6, x) = C(0.6)°(1 0.6) - ) = b(*), say.
Accordingly, the probability that four of five heaters work
satisfactorily is
b(4) = °C,(0.6)(0.4) = 0.2592.
Next, we note that atmost two of five heaters do not work
satisfactorily if at least three of them work satisfactorily. The
probabilityof this to happen is
b(x> 3) = b(3) + b(4) +
b(5)
Co6'0.4. c 0.6'0.4)+ co.6
=0.3456 + 0.2592 + 0 07776 = 0.68256.
430
TU. Engineering Mathematiey
Example 10. Thvo persons Aand Bplay agame in
chances of winning are in the ratio 3:2
winning at least three games out of six games played
nhich
Find A's chanothoy

» From what is given, we note that the probability that ,

game is p= 3/5.
Since six games are played, the probability function
Successes for A in these 6 games is given by
the gving
probability function binoma
6-x
3
5
= b(x) say

Hence, the probability of A winning at least three of siz mo

is
b(x> 3) = 1- br< 3) = 1-b(0) + b(1) +b(2)
6 5 2

--()E) 2+
2 2+ (6x 3x 2) + 15 x3= 1-x
= 1 175
5 5

= 1-0.1792 = 0.8208.
Thus,A has about 82% chance of winning at least three of the
six games.

Example 11. An airline knows that 5 percent of the peopi

making reservations on a certain flight will not turn up.
Consequently, their policy isto sell52 tickets for a flight that cat
there willk
only hold 50, passengers. What is the probability that
aseat available for every pa ssenger wvho turns upí
>» The probabilitythat a passenger will not turn up is p =
number
I
Therefore, the probability that, of the 52 passengers, function
of passengers will not turn up is given by the probability
b(52, 0.05, x) = ( CNO.05)0.95,2
= b(), say.
2ProbabilityDistributions 431

There willbe a seat for every passenger who turns up if x> 2.

happen is
The probability of thisto
b(x >2) = 1- b(x< 2) = 1-b(0) + b(1)}
= 1-(0.95)°2 + 52(0.05)(0.95,
= 1- 0.2595 = 0.7405

Thus, the probability that there will be a seat available for

every passenger who turns up is about 74%.

Example 12. In a large number of parts manufactured by a

machine, the mean number of defectives in a sample of 20 is 2.Out
of 1000 such samples, howmany would be expected to contain at
least 3 defective parts?
» The probability that a part is defective is, from what is given,
p= 2/20= 1/10. Therefore, the probability function for the
distribution is

1
b20, = bx), say
Therefore, the probability that there are at least 3 defective
parts in a sample of 20 items is
b(* 2 3) = 1- b(x <3) = 1- b (0) b
(1) b(2)
20 19 2 18

+ +
20C 2
= 1-(0.1216 + 0.2702 + 0.2852) = 0.323.
Hence, out 1000 samples the expected number of samples
that contain atof least three defective parts is 0.323 x 1000 = 323.
Example 13. In 800 families with five children each, how many
funilies
and 3 would be expected to have (i) 3boys and 2girls, (i) 2boys
girls, (iii)no girls, (iv) at most 2 girls, by assuming
probabi
»
lities of births of boys and girls to be equal.
The probability that a 1/2. Therefore, the
probability that there child is a girl is p=
are xnumber of girls out of five children (in
a
family) is given by the binomial probability function
 TU. Engineering Mathematies.Iy
432
5 -x
1 1
1
b| 5, 2* 2 = b(x) say.

Therefore, the probability that there are two girls among five
children (in a family) is
5x 4 1 5
b(2) = = 0.3125.
2 g5
Consequently, out of 800 families with
with
5 children
each theis
expected number of families two girls
800 x 0.3125 = 250. Since probabilities of girls and boys are
equal, the expected number families having
of two boys among
five children is also 250.
Next, the probability that there is no girl among five children is
5
1
b(0) ==c, 32
= 0.03125.

Accordingly, out of 800 families with 5 children each. the

expected number of families with no girl child ie
800 x 0.03125 = 25.
Lastly, the probability that there are at most two girls among
five children is
b(xs 2)= b0) + b(1) + b(2)
5

1 1
= (1+ 5+ 10)
g5 2

Accordingly, the expected number of families being considered

1
with at most two girl children is 800 x = 400.
2
Exercises
of a binomial distribution are 4 and 4/3
1. The mean and variance
respectively. Find P(r2 1).
exactly 3
2. Six coins are tossed. Find the proability of getting (i)
heads (ii) at least 3 heads (iii)at least one head.
result
3. Acoin is tossed6 times. Find the probability that the tosses
in (i) heads only (ii) 3heads and 3tails (iii) 5 or more heads.
4. Acricket team has 2/3 chance of winning a match. Ifit plays 4
matches, find the probability that it wins (i) 2 matches (ii) at least
one match.
438 T.U. Engineering Mathematies.Iy
5. In a consignment of electric lamps, 5% are defective
sample of 8 lamps is taken for inspection. What. N
that it has one or more defectives?
is the rprobaabniomlity
6. The incidence of occupational disease in an industry is
workers have 20% chance of suffering from it. such that
What
probability that out of six workers four or more will is the
from the disease.
be suffering
7. Afleet of 20 aeroplanes are sent on an operational
chance that an aeroplane fails to return from the flight is 5 flight. The
Find the probability that one plane does not return. percent.
8. The probability that a bomb dropped from a plane will. hit its
is 0.2. If six bombs are dropped, find the probability that (i) target
two will hit the target (i) atleast two will hit the target. exactly
9. It wasfound that 10% of boys in a certain class were suffering fron
short sight. What is the probability that a random sample of 5boys
willcontain (i)no boy suffering from short sight, (i) exactly 1bÍy
suffering from short sight, (iii) not more than 4 boys suffering from
short sight ?
10. A product is 0.5% defective and is packed in cartons of 100. What
percentage of cartons contains not more than 3 defectives?
11. If the probability of success is 0.01, how many trails are necessary in
order that the probability of at least one success is 0.5 or more?
12. Fit a binomial distribution for each of the following frequency
distributions. Also, ind the corresponding theoretical estimates or

(i) I,: 1 2 3 4

5 29 36 25 5

9 10
0 1 7
(ii);: 2 3 4 5 6

: 6 20 28 12 8 6 0 0

Answers
0.1094
1. 0.9986 2. 0.3125; 0.65625: 0.984375 3. 0.0156; 0.3125;
4. 0.2963; 0.9877 5. 0.3366 6. 0.0169
2 ProbabilityDistributions 439

7. 0.3585
8. 0.246, 0.345 9. 0.5905; 0.3281;0.9996
11. 69
10. 99.83

(0.49 (0.51)4- x
12. (Ö) P(r) = (C)
EO = 6.76, F(1) = 26, F(2) = 37.47, F(3) = 24, F(4) = 5.76
-*
(ü) Pr) = ("CX0.2175 (0.7825)10
R(O) = 6.9, F(1) = 19.1, F(2) = 24, F(3) = 17.8, F(4) = 8.6
F(5) = 2.9, F(6) = 0.7, F(7) = 0.1, F(8) = F(9) = F(10) = 0.

7.3.2 Poisson Distribution

Let us recall that the binomial probability function b(n, p, x) is
given by the formula
b(n, p, *) = "C, p (1- p- , x = 0, 1,.... n...(1)
and that the mean of the binomial distribution is u = np.
If n is very large and p is very small but u is finite, we will
prove below that
eu
b(n, p, x) ..(2)
x!
Proof. We have to show that the formula (2) is valid when x is
zero or a (any) positive integer. We employ the method of
induction.

From expression (1), wefind that

b(n, p, 1) = np (1- p) b(n, p, 0)
= H(1+ p+p"t) b(r, p, 0) bn, p, 0)
ue

This shows that the formula (2) is valid for x= 1.

we suppose that the formula (2) is valid for x = k, where k is
Some positive integer. That is,
eu (3)
b(n, p, k) =
k!
440 T.U. Egineering Mathematiesty
Then the relation (1) gives
b (n, p, k+ 1) n- k
b (n, p, k) k + 1 1- p

=
- kp (1- p)= u- kp)
k + 1 k+ 1 (+p+p+,
(for small p )
k + 1

so that b (n, p, k+ 1) = k+ 1 k! using (3)

k+ 1
e

(k + l)!
This shows that the formula (1)is valid for x = k+ lae w.
Hence, by induction, we infer that the formula (2) is valid for
every positive integral value of x.
For x = 0, we have

b(n, p, x) = b(n, p, 0) = (1- p n

so that

log b(n,p, 0) = n log

[-)
= n,
2n2 3n
(on using the logarithmic series
3
= -
3,2

If n is very large, p is very small and u is finite, the terns

u'/n= Mp, may be ignored, and we get
log b(n, p, 0) - u. Hence
b (n, p, 0) = e
This result shows that the formula (2) is valid for x= 0.
or
positive integerr
Thus, the formula (2) is valid when x is any
Zero.
ZProbabilityDistributions
441

Pojsson Probability Function

e
value for b (n, p, x) given by
The approximate x!
formula (2), which is valid when n is very large and p is very
small but u = np is finite, is called the Poisson Probability
Function, named after the French mathematician S.D. Poisson
1781-1840). We denote it by p (u, x). Thus,

p (u, x) = x = 0, 1, 2, ... (4)

!

The probability distribution for which this function is the PD.F.

is called the Poisson distribution and u is called theparameter
of the distribution. Evidently, this distribution is an infinite
descrete probability distribution. The random variable x
associated with a Poisson distribution is called a Poisson variate.
The table of Poisson distribution is as given below:

0 1 2 3

|P(4,z) : le
3!

It is to be noted that we have introduced the PoiSson

astrbution as an approximation (limiting case) of the binomial
distribution. According to this approximation, the probability
function P(4, *) gives the probability of exactly xsuccesses in
nindependent trials when (i) nis very large,(ii) each trial yields
a success or failure, (iii) the probabilityVp of success in each trial
1s the same but is very Small, and (iv) the mean of the
distribution, np, is finite.
The Poisson distribution can be studied independently of the
binomialIdistribution. In such studies, the formula (4) is taken as
the definition of the P.D.E. for the Poisson distribution without
making any reference to the approximation made through the
formula (2).
442 TU. Engineering Mathematiey
We easily verify that p(4, ) is indeed a

function in the sense that probability dor

() p (4, x,) 2 0 for x, = 0, 1, 2,
and (ii) Ep(4, x,)= 1.
When there is no ambiguity, we write p(p, x) as just
p(r)
Mean
The mean of the binomial distribution is u- .
Poisson distribution u continues to be finite and it servesS as the
mean.

Variance and Standard deviation

The variance of the binomial distribution is

= V= np(1- p) = l-p) = -4p

For the Poisson distribution, uis finite andp is small; in th
case we get

Thus,for the Poisson distribution,

Mean = Variance = u.
and Standard deviation= G= Vu

Example 1. The probabilities of a Poisson variate taking

values 3 and 4 are equal. Calculate the probabilities of the var?
taking the values Oand 1.

We are given that p(u, 3) = p(u, 4). That is,

e e
3! 4!
4!
This gives 3!
= 4.

Consequently,
Z.ProbabilityDistributions 443

e e4
P(u, x) =
x! x!

isthe probability density function for the given variate x. We

find that
-4
-4
P(H, 0) = e = 0.01832, and p(H, 1) = 4
1! = 0.0733.

These are theprobabilities of the variate x taking the values0

and 1(respectively).

Example 2. If x is Poisson Uariate such that

P(r =1)= 0.2x P(x= 2), find P(x = 0).
» For the Poisson distribution, the probability density function is
e
p(u, x) = ) I= 0, 1, 2,
x!
Here, it is given that
P(H, 1)= 0.2 x p(u, 2)
eu' = 0.2 x e
i.e.
1! 2!
or 1=
02x
2
This gives u = 10. Therefore,
e
P(x = 0) = p(u, 0) = =e
0

= e0= 4.54 x 10 -

Example 3. Acar hire firm has 2cars, which it hires out day by
day. The demand for a car on each day is distributed as a Poisson
distribution
certain day (i)
with mean 1.5. Calculate the probability that on a
neither car is used, and (ii) some demand is refused.
»
Here, H = 15 and the
Poisson probability function is
1.5
p(1.5, x) = (1.5)*
= p () (say)
Then the probability that neither car is used is
444
T.U. Engineering Mathematiey
p(0) = e - 1.5
0.2231
Also, the probability that some demand is refused =
that there are more than 2 demands
= P( x > 2)
= l- (p (xs 2) ) =
l-
probail ty
p(0) + p (l) + p (2
1.5 1.5 1.5
1-e + e
(1.5) + 2!
(1.5) 2

- 1.5
= 1- e (3.625) = 1- (0.2231) x 3.625
= 0.1913

Example 4. Alpha particles are emitted by a

at an average rate of5 in a 20 minute interval. radioactive
Using
SOur%
distribution, find the probability that there will be (i)Poistson
emissions, and (ii) at least two emissions, in a 20 minute internal
Here, u= 5 and the probability function is
p(5, x) =
e55*
p(«) (say)
Then the probability that there will be two emissions is

P(2) = = 0.0842
2!
Also, the probability that there will be at least two emissions i
p(r> 2) =1- p(x< 2)
= l- p(0) + p(1)
= 1- le+ 5e-5]
= 1- 6e -5 = 0.9596.

by afirm
Example 5. Given that 2% of the fuses manufacturedprobability
are defective, find, by uSing Poisson distribution, the
that a box containing 200 fuses has (i) at least one defectivefus
(ii) 3or more defective fuses.
2/100
The probability that a fuse is defective is given as p: =
= 0.02. Also, n= 200, so that u = np = 4.
7ProbabilityDistributions 445

Hencethe Poisson probability function is

e -4 4
p(4,x) = = p (x). say.
x!

Therefore:
Probabilitythat the box contains at least one defective fuse is
(i) Pr
p(x >1) = 1-p(x< 1) = 1- p(0)
= 1-p -4 = 0.982

(G) Probability that thebox contains 3or more defective fuses is

p(x 2 3) = 1-p (x<3)
=1- [p(0) + p(1) + p (2))
-4
e + e -4 4+ -442|
-1 2!
-4
= 1- (1+4+8) e = 0.762.

Example 6. The probability that an individual suffers a bad

reaction from a certain injection is 0.002. Using Poisson
distribution, determine the probability that out of 1000
individuals, (a) exactly 3, (6) more than 2will suffer a bad
reaction.
» The probability that the individual suffers a bad reaction is
gven as p= 0.002. Also. n= 1000, Therefore, u = np = 2.
The corresponding Poisson probability function is
e
p(2,x) = x! p(x), say
Therefore, the probability that exactly 3 will suffer a bad
reaction is
p3) = 4 2
3!
x 0.13534 = 0.1804.

Also, the probabilitythat more than 2 will suffer abad reaction

P(* > 2) = 1- p(x s 2) = 1-p(0) +

p(1) +p(2)
446 TU. Engineering Mathematie
p-2,90 2
2 e -2
= 1 0! 1!
+
2!
2
= 1- (1+ 2+ 2) e
- 1- 0.67667
= 1- 5 x 0.13534!= -
0.32333
Example 7. In a certain factory turning out razor
is asmallchance of 0.002 for any blade to be
supplied in pachets of 10.
defect
Using Poission ive.blade
Thes, tha
blas
are

the approximate number of packets containing () no iM

blade, (i) one defective blade, and (iii) two defective defecth
distribution,
consignment of 10,000 packets.
in blades,
>» The probabilitythat a blade is defective is given as p= 0.
Also, the number of blades in a packet is n = 10. Therefore
u = np = 10x 0.002= 0.02
Hence the Poisson probabilityfunction is

p(0.02, x) =
e
-0.02(0.02
x!
= p(*), say

Therefore,
(i) Probability that apacket contains no defective blade is
0.02
x (0.02) =
o-0.02
e
Po = p0) = 0!
0.9802.
Consequently, the number of packets containing no detet
blades is 10,000 x P, = 9802.
(i) Probability that a packet contains one defective blade s
-0.02 x(0.02)!
P, = p(l) = 1!
0.9802x 0.02 0.019604
defecti:
one
Consequently, the number of packets containing
blade (each) is 10000 x p, 196.
bladesis
defective
(iiüi) Probability that a packet contains two
z ProbabilityDistributions
447
0.02
x (0.02)2
P, = p(2) = 2!
1
)
2 (0.9802 x 0.004) 0.00019604
Consequently the number of packets containing two defective
blades (each) is 10,000 x P2 2.

Example 8. While forming 36-digit binary numbers a

malfunctioning digital computer has found to form a digit
incorrectly about 1times in 1000. Assuming the errors in forming
diferent digits are independent, find the probability of having ()
zero, (ii)one, (iii) more than one, and (iv) less than five incorrect
digits in a given 36-digit number.
The probability that a digit is incorrectly formed is given as
1/1000 = 0.001. Also, the number of digits in a number is
n= 36. Therefore, the mean number of incorrectly formed digits

H= np = 36 x 0.001 = 0.036.
Hence the probability function for the distribution of incorrect
digits is the Poisson probability function
-0.036
p(0.036, x) =
e
(0.036)*

p(x) (say).
Therefore, in a given 36-digit number,
(i) the
probability having no incorrect digit is
(ii) the
P(0) = e

-0.036(0.036)"
0!
-0.036
= 0.96446.

probability of having one incorrect digit is

e0.036
P(l) = X 0.036
= 0.0347.
1
(ii) the
probability of having
more than one incorrect digit is
P(*> 1) = 1- plx s l) = 1- p(0)-
P()
(

448
TU. Engineering Mathemati
1- 0.9646 0.0347 =
(iv) the probability of having less than
0.0007
five
5) = p(0) + p(1) + p(2) + ip(3)ncorrect
+ ddigts i
0.036
p(4)
= 0.9646 + 0,0347 + x(0.0362
2!
0.036
x(0.036)' 0.036
3!
4
= 0.9999.

Example 9. Fit a Poisson distribution tothe

distribution:* fol owing fequenes
1 2 3 4

122 60 15 2 1

Find the corresponding theoretical estimates for f.

» Here,
Ef. = 122 + 60 + 15 + 2.+ 1= 200.
Ex,f. = (0x 122) + (1 x 60) + (2 x 15) + (3 x 2) +(4x 1)= 100.
Therefore, the mean of th given frequrency distribution is

* As indicated under Example 17 of Section 7.3.1, the meari ot

frequency distribution is

X =

Fitting a Poisson distribution to this frequency distribution consss

Sothatthe
in finding a Poisson distribution whose mean is = X,
P.D.F. is
Pr) = x= 0, 1, 2,
r!
frequences
Then Fx,) = (2f)Pu) arecalled the theoretical
funderthe
which serve as thetheoretical estimates of the given
fitting.
7.ProbabilityDistributions
49

100
= 0.5
200

Accordingly, the Poisson distribution that fits the given

distribution has the mean equal to 0.5, so that the PD.F is
-0.5
(0.5)
P(x) =
x!

The corresponding theoretical frequencies are

F(x) =(2) P,), x,= 0, 1, 2, 3, 4,;
that is,
F(0)= 200 x e -0.5 = 200 x 0.6065 = 121.3,

F(1) = 200 x e -0.5 x 0.5 = 60.65,

- 0.5 1
F(2) = 200x e ;(0.5) = 15.1625,
-0.5
F(3) = 200 x e (0.5)° = 2.527,
F(4) = 200 × e -0.5 x(0.5) -0.3159.
4!

These are the theoretical estimates of the given f fo, fa

f¡o f; respectively, accordingly to the Poisson distribution fitting.

Exercises
L. In a Poisson distribution, if p (2) = (2/3) p(1), find p(0), P(3)
and p(x > 3).
the number of accidents occuring on a highway each day is a
Poisson variate with equal to 3, what is the probability that
no accidents occur on mean
arandomly chosen day.
3. Supp08e 2% of
the probability the items produced by a machine are defective. Fnd
that there are 3 defective items in asample of 100
items.
4.
Amanuf
Poisson acturer produces
100ehipsdistribution, find
no defectives
IC chips, 1% of which are defective. Using
the probability that in abox contaning
are found.
450
nu Engineering Mathemate,
5. The number of accidents in a year to auto-drivera
Poisson variate with mean equal to 3. Out of 1000
approximately the number of drivers with () such:
no
than 3 accidents, in a year. accident.
telephone calls
6. The average number of per
exchange between 10 a.m. and 12 noon is 4. Find minnte
booked
selected day 2 or more calls the
that on a randomly
booked between 10 a.m. and 12 noon.
On how per
many miofnutae
prohat
would you expect booking of 2 or
more calls per minute days
time gap? dun gh
7. Suppose 300 misprints are distributed randomly throughout ab
e
of 500 pages. Find the probability that a given page contans
exactly 2 misprints, (ii) 2 or more misprints.
8. Acertain screw making machine produces on an average 2 defecti
screws out of 100, and packs them in boxes of 500. Find the
screws.
probability that a box contains 15 defective
fuses will be foundi.
9. Find the probability that atmost 5 defectivepercent of such fuses an
box of 200fuses ifexperience shows that 2
defective.
percent of items of his productu
10. A manufacturer knows that one
100, what is the
defective. If he sells the items in boxes of
contain three or mor
probability that a box picked at random will
defective items?
following frequeng
11. Fit a Poisson distribution for each of the theoretical estimats
distributions. Determine the corresponding
for f;
2 4
1
(i)
19 5 0
211 90

1 3 4
(ii)
38 22 9 1
46

3 4 5
(iii) x, : 2

173 168 37 18

Answers
2. 005
I. p(0) 0.2636, pb) 0.1041, p(r> 3)= 0. 1506
Z.Probability Distributions 451

3. 0.180 4. 0.3679 5. (i) 50 (ii) 353

6. 0.9084: 332 7. 0.1; 0.122 8. 0.035

9. 0.785 10. 0.08

0.44
(0.44)
11. ) Px) = x!
F0) = 209.3, F(1)= 92.09, F(2) = 20.23, F(3)= 2.972,
F(4) = 0.327
0.974
e (0.974)
(ii) P(x)= x!

F(0) = 43.8, F(1) = 42.66 F(2) = 20.78, F(3) = 6.75

F(4) = 1.64
-0.7825
(0.7825
(iii) P(x) = x!
F(0) = 182.9, F(1) = 143.12, F(2) = 55.72; F(3)= 14.61,
F(4) = 2.86; F(5) = 0.45.