Documente Academic
Documente Profesional
Documente Cultură
Theoretical distributions
In this chapter, we shall consider some important discrete and continuous probability distributions,
which occur frequently in many engineering and other applications.
Binomial distribution is derived from an experiment known as Bernoulli trial named after the swiss
mathematician James Beroulli.
Definition 2.1.1 A random variable X is said to follow a binomial distribution with parameter n and p
if its probability mass function is given by
nC x q n−x , x = 0, 1, 2, · · · , n
xp
P (X = x) =
0, otherwise
Note 1
1. The parameter p is the probability of success in a single trial and n is the number of trials. n is
also known as the index or the degree of the binomial distribution.
3. Binomial distribution is used in games of chance, military defense strategies, failure analysis and
many other situations.
2. p+q=1
1
2.2 Statistics and Queueing Theory
5. Mode of the binomial distribution is that value of x for which p(x) is maximum.
n
X
tX
MX (t) = E(e ) = etx P (X = x)
x=0
n
X
= etx nCx px q n−x
x=0
n
X
t
= e nCx (pet )x q n−x
x=0
= [pe + q]nt
0
First Moment = Mean = E(X) = MX (0)
0
MX (t) = n(q + pet )n−1 pet
0
MX (0) = np
00
Second Moment = E(X 2 ) = MX (0)
00
MX (t) = np[et (q + pet )n−1 + et (n − 1)(q + pet )n−2 pet ]
00
MX (0) = np[1 + (n − 1)p] = np + n2 p2 − np2
= np + n2 p2 − np2 − n2 p2 = np − np2
= np(1 − p) = npq
Problem 1 The mean of a binomial distribution is 20 and standard deviation is 4. Find the parameters
of the distribution.
Given np = 20 − − − −(1)
√
and npq = 4
⇒ npq = 16 − − − −(2)
Using (1) in (2), we get
16 4
(2) ⇒ 20q = 16 ⇒ q = =
20 5
4 1
∴ p=1−q =1− = and
5 5
1
∴ (1) ⇒ n. = 20 ⇒ n = 100
5
1
∴ the parameters are 100, .
5
Problem 2 Comment on the following ”The mean of a binomial distribution is 3 and variance is 4”.
Given np = 3 − − − −(1)
and npq = 4 − − − −(2)
Using (1) in (2), we get
4
(2) ⇒ 3q = 4 ⇒ q = > 1, which is not true, since the probability value can not be greater than 1.
3
So, there is no binomial distribution with this data, the statement is false.
Problem 3 Six dice are thrown 729 times. How many times do you expect atleast 3 dice to show a five
or a six?
Solution: Success is getting 5 or 6 in a die. Let X denote the number of success when 6 dice are thrown.
∴ X is a binomial random variable with parameter (n, p).
∴ P (X = x) =n Cx px q n−x , x = 0, 1, 2, · · · , n
2 1
Given n = 6 and p =probability of getting 5 or 6 = =
6 3
1 2
∴ q =1−p=1− =
3 3
x 6−x
1 2
∴ P (X = x) =6 Cx , x = 0, 1, 2, · · · , 6
3 3
p(X ≥ 3) = 1 − P (X < 3)
= 1 − {P (X = 0 + P (X = 1) + P (X = 2))}
( 0 6−0 1 6−1 2 6−2 )
6C 1 2 6 1 2 6 1 2
=1− 0 + C1 + C2
3 3 3 3 3 3
1
=1− {64 + 192 + 240}
36
496 233
=1− =
729 729
233
when 6 dices are thrown 729 times, the number of times atleast 3 dice show 5 or 6 is 729 × = 233.
729
Problem 4 In 256 sets of 12 tosses of a fair coin,in how many cases may one expect 8 heads and 4 tails?
1
Given n = 12 and p =probability of getting head=
2
1 1
∴ q =1−p=1− =
2 2
x 12−x 12
12 1 1 12 1
∴ P (X = x) = Cx = Cx , x = 0, 1, 2, · · · , 12
2 2 2
12
12 1
∴ P (X = 8) = C8 = 0.12
2
∴ the number of times to get 8 heads and 4 tails appearing in 256 sets= 256 × 0.12 = 30.72 ≈ 31times.
Problem 5 Ten coins are thrown simultaneously. Find the probability of getting atleast 7 heads?
Solution: Here n=10. Let X denotes number of heads. ∴ X follows a binomial distribution.
∴ P (X = x) =n Cx px q n−x , x = 0, 1, 2, · · · , n
1 1
Now p=probability of getting a head= . ∴ q = 1 − p =
x n−x 2 2
10 1 1 10 1 10
∴ P (X = x) = Cx = P (X = x) = Cx , x = 0, 1, 2, · · · , 10
2 2 2
Required P (X ≥ 7) = P (X = 7) + P (X = 8) + P (X = 9) + P (X = 10)
10 10 10 10
10 1 10 1 10 1 10 1
= C7 + C8 + C9 + C10
2 2 2 2
10
1
= [120 + 45 + 10 + 1] = 0.1718
2
5
1 3 t
Problem 6 A discrete random variable X has moment generating function MX (t) = + e .
4 4
Find E [X] , V ar [X] and P (X = 2).
n
Solution: We know that moment generating function MX (t) = q + pet − − − − − (1)
1 3 t 5
Given MX (t) = + e . − − − − − −(2)
4 4
By uniqueness of MGF, (2) is the moment generating function of binomial distribution.
3 1
∴ n = 5, p = , q = .
4 4
3 15
∴ E [X] = np = 5. =
4 4
3 1 15
and V ar [X] = npq = 5. . =
4 4 16
5−2 2 3
32
1 3 1 45
∴ P (X = 2) =5 C2 5
= C2 = = 0.0879
4 4 4 4 512
Exercise
1. The mean and variance of a binomial distribution are 8 and 6. Find P (X ≥ 2).
2. The mean and s.d.(standard deviation) of a binomial distribution are 5 and 2 respectively. Determine
the distribution.
3. Out of 800 families with 4 children each, how many families would be expected to have (1) 2 boys
and 2 girls (2) atleast one boy (3) at most 2 girls (4) children of both sexes.
4. Out of 2000 families with 4 children each, how many would you expect to have (1) at least 1 boy
(2) 2 boys (3) 1 or 2 girls (4) no girls?
Poisson distribution is a limiting case of binomial distribution when the probability of success p is very
small and number of Bernoulli trials n is very large so that np = λ is a constant of moderate value. Thus
poisson distribution is also associated with Bernoulli trials and it relates to rare events. It was discovered
by the French mathematician Denis Poisson when he applied probability theory to law suits, criminal
trials, etc.
Definition 2.2.1 A random variable X is said to follow a Poisson distribution with parameter λ > 0, if
its probability mass function is given by
λx
P (X = x) = e−λ x = 0, 1, 2, · · · , ∞
x!
Note 3
1. Mean=Variance= λ.
t −1)
3. Moment generating function of Poisson distribution is MX (t) = eλ(e .
∞
X
MX (t) = E(etX ) = etx P (X = x)
x=0
∞
X e−λ λx
= etx
x!
x=0
∞
X (λet )x
= e−λ
x!
x=0
−λ λet t −1)
= e e = eλ(e
t −1)
∴ MX (t) = eλ(e
t −1)
MX (t) = eλ(e
0
First Moment = Mean = E(X) = MX (0)
0 t −1)
MX (t) = λet eλ(e
0
MX (0) = λ
00
Second Moment = E(X 2 ) = MX (0)
00 t −1) t −1)
MX (t) = (λet )2 eλ(e + λet eλ(e
00
MX (0) = λ2 + λ
= λ2 + λ − λ2 = λ
Problem 1 six coins were tossed 6400 times using the Poisson distribution. Find the probability of
getting 6 heads 10 times.
1
Solution:Probability of getting one head with one coin=
2
6
1 1
∴ Probability of getting six heads with six coin= =
2 64
1
Here n=6400. ∴ λ = np = 6400 × = 100.
64
10010
P(getting 6 heads 10 times)= P (X = 10) = e−100
10!
Problem 2 A car hire firm has two cars which is been hired for day by day basis with mean of 1.5. Use
Poisson distribution to find on a particular day.
2. no car is used.
3. no demand is refused.
= 1 − [P (X = 0) + P (X = 1) + P (X = 2)]
(1.5)0
2. P (X = 0) = e−1.5
0!
= e−1.5 = 0.223
3. P (X ≤ 2) = [P (X = 0) + P (X = 1) + P (X = 2)]
Exercise
1. A manufacturer of pins know that 2% of his products are defective. If he sells pins in boxes of 100
and guarantees that not more than 4 pins will be defective. What is the probability that a box will
fail to meet the guaranteed quality.
3. It is known that probability that an item produced by a certain machine will be defective is 0.01.
By applying Poisson approximation, show that the probability that a random sample of 100 items
2
selected at random from the total output will contain not more than one defective item is
e
t −1)
4. The MGF of a random variable X is e3(e , find P (X = 1).
A random variable X is said to follow Geometric distribution, if it assumes only non-negative values and
its probability mass function is given by
P (X = x) = q x p, x = 0, 1, 2, . . . , 0 < p < 1
where q = 1 − p.
where q = 1 − p.
∞
X
MX (t) = E(etX ) = etx P (X = x)
x=0
∞
X
= etx q x p
x=0
∞
X
= (qet )x p
x=0
= p[1 + qet + (qet )x + (qet )x + . . . ]
p
∴ MX (t) = (1−qet )
0
First Moment = Mean = E(X) = MX (0)
0
MX (t) = (−1)p(1 − qet )−2 (−qet )
0 pq q
MX (0) = 2
=
p p
00
Second Moment = E(X 2 ) = MX (0)
00
MX (t) = pq[(−2)(1 − qet )−3 (−q(et )2 ) + (et )(1 − qet )−2 ]
00 q(1 + q)
MX (0) =
p2
∞
X
P (X > r) = qxp
x=r
= q r p[1 + q + q 2 + . . . ] (1)
r 1
= q p = qr (2)
1−q
Now
P (X > s + k, X > s)
P (X > s + k/X > s) =
P (X > s)
q (s+k)
= = q k = P (X > k) (3)
qs
A continuous random variable X is said to follow an exponential distribution with parameter λ > 0 if
its probability density function is given by
λe−λx ,
when , x > 0
f (x) =
0, otherwise
Z∞
tX
MX (t) = E(e ) = etx f (x)dx
x=0
Z∞
= etx λe−λx dx
0
Z∞
= λ e−(λ−t)x dx
0
" #
e−(λ−t)x
= λ
−(λ − t)
λ
= = λ(λ − t)−1
λ−t
0
First Moment = Mean = E(X) = MX (0)
0
MX (t) = λ(−1)(λ − t)−2 (−1)
0 λ 1
MX (0) = =
λ2 λ
00
Second Moment = E(X 2 ) = MX (0)
00
MX (t) = λ(−2)(λ − t)−3 (−1)]
00 2λ 2
MX (0) = 3
= 2
λ λ
Z∞
P (X > r) = λe−λx dx
x=r
∞
e−λx
= λ = e−λr (4)
λ r
Now
P (X > s + k, X > s)
P (X > s + k/X > s) =
P (X > s)
e−λ(s+k)
= = e−λk = P (X > k) (5)
e−λs
Normal distribution is the most important continuous probability distribution in statistics both from
practical and theoretical point of views.
Definition 2.5.1 A continuous random variable X with parameters µ and σ 2 is said to follow a normal
distribution if its probability density function is given by
!2
x−µ
1 −
f (x) = √ e σ , −∞ < x < ∞, σ > 0
σ 2π
Note 4
1. The notation X N (µ, σ 2 ) means X is normally distributed with mean µ and variance σ 2
1
2. In the binomial distribution with parameters n and p, when n is very large and p is nearly the
2
binomial approaches normal.
Rx2
3. Let X N (µ, σ 2 ). Then P (x1 < X < x2 ) = f (x)dx. For any x1 , x2 this integral cannot be
x1
evaluated in closed form and so it should be evaluated by numerical methods. As µ and σ vary, we
have different function and every time we have to make different computations. So it is convenient
to have a fixed table which could be used for all such computations. This is done by transforming
X −µ
X into Z = .
σ
1. P (X ≥ 19)
3. K if P (X > k) = 0.24
z = 0.5 − 0.3413
0 0.3413 1 ?
0.5
= 0.1587
12.5 − 16
2. When X = 12.5, Z = = 1.16
3
19 − 16
When X = 19, Z = =1
3
∴ P (12.5 < X < 19) = P (−1.16 < Z < 1)
= P (−1.16 < Z < 0) + P (0 < z < 1)
= P (0 < Z < 1.16) + P (0 < z < 1) 0.3710
= 0.3770 + 0.3413 z
-1.16 0.3710 0 0.3413 1 1.16
= 0.7183
3. P (X > K) = 0.24
k − 16
When X = K, Z = = z1
3
∴ P (X > K) = 0.24
⇒ P (Z > z1 ) = 0.24
⇒ P (0 < Z < z1 ) = 0.5 − 0.24 = 0.26
∴ the z1 is the value of Z corresponding to the area 0.26 and z1 = 0.7 (from normal table)
K − 16 K − 16
∴ = z1 ⇒ = 0.7 ⇒ K = 3(0.7) + 16 ⇒ K = 18.1
3 3
Problem 2 In a test on 2000 electric bulbs, it was found that the life of a particular make was normally
distributed with an average life of 2040 hours and standard deviation of 60 hours. Estimate the number
of bulbs likely to burn for
0.5
= 0.1587
bulbs expected to burn for ≥ 2100 hours = 2000 × 0.1587 ≈ 159.
1950 − 2040
2. When X = 1950, Z = = −1.5
60
∴ P (X ≤ 1950) = P (Z ≤ −1.5)
= 0.5 − P (0 < z < 1.5)
z = 0.5 − 0.4332
? -1.5 0 0.3413 1.5 ?
0.5
= 0.0668
∴ the number of bulbs expected to burn for less than 1950 hours = 2000 × 0.0668 ≈ 134.
1920 − 2040
3. When X = 1920, Z = = −2
60
2160 − 2040
When X = 2160, Z = =2
60
∴ P (1920 < X < 2160) = P (−2 < Z < 2)
= P (−2 < Z < 0) + P (0 < z < 2)
= P (0 < z < 2) + P (0 < z < 2)
0.4772
= 2P (0 < z < 2)
z
-2 0 2
= 2(0.4772) ?
= 0.9544
∴ the number of bulbs expected to burn for 1920 hours to 2160 = 2000 × 0.9544 ≈ 1908.
Problem 3 In a normal distribution 31% of the items are under 45 and 8% are over 64. Find the mean
and s.d. of the distribution.
X −µ
Solution: Let X follows N (µ, σ 2 ), Z =
σ
31 8
Given P (X < 45) = = 0.31 and P (X > 64) = = 0.08
100 100
45 − µ
When X = 45, Z = = −z1 ,
σ
31
∵ P (X < 45) = = 0.31
100
0.19
⇒ 45 − µ = −σz1
0.31 ∴ z1 is the value of z corresponding to the area
−z1
z Rz1
z=0
ϕ(z)dz = 0.19
0
∴ z1 = 0.495(from tables)
∴ 45 − µ = −0.495σ − − − − − (1)
64 − µ
When X = 64, Z = = +z2 , ∵ P (X < 64) = 0.08
σ
⇒ 64 − µ = +σz2
0.42
∴ z2 is the value of z corresponding to the area
0.08 Rz2
z ϕ(z)dz = 0.42
z=0 Z2
0
∴ z2 = 1.405(from tables)
∴ 64 − µ = 1.405σ − − − − − (2)
Solving (1) and (2), we get µ = 49.95, σ = 10 or µ = 50, σ = 10
Exercise
1. If X is normally distributed with mean 8 and s.d 4. Find
(a) P (5 ≤ X ≤ 10)
(c) P (X ≥ 15)
2. In a distribution exactly normal 7% of the items are under 35 and 89% are under 63. What are the
mean and standard deviation of the distribution?
3. The marks of 100 students in an exam are found to be normally distributed with mean 70 and S.D.
5. Estimate the number of students whose marks will be