stoc proc kg
TRANSCRIPT
-
8/10/2019 Stoc Proc KG
1/18
Author: Dr. K. Gururajan
Assistant Professor
Department of Mathematics
Malnad College of Engineering, Hassan573 201
UNIT 8:
Jointly Distributed Random Variables
Introduction: So far the study was restricted to one dimensional random variable,distribution function, and other related aspects. However, in real world, one come
across a number of situations where we do find the presence of two or correlated random
variables. For example, consider the experiment of finding how far e - learningprogramme initiated by VTU Belgaum has become popular among the engineering
students? To find this say, authorities collect feed back from the students by visiting their
institutions. Let the problem be about finding the opinion of students regarding two
parameters; (i) quality of transmission from studio situated at Bangalore which we call it
as X and (ii) students interest in this kind of programs which we shall refer it to as Y.For convenience of discussion, authorities visit seven colleges located at different parts of
the state and results are in the following table. We assume that these are given in terms ofpercentages.
Engg.
Colleges
PGA SSIT BVB GSSIT AIT SBMJCE KLE
X X1 x2 x3 x4 x5 x6 x7
Y y1 y2 y3 y4 y5 y6 y7
In problems like this, we/authorities are certainly interested to learn the mood of the
students/teachers about the e learning programme initiated by us of course with hugecost. It is known to you that one satellite channel has been completely dedicated for this
purpose in India. Many people are involved in this programme to reach the unreached
and needy.
One comes across many illustrative examples like that. Therefore, there is a necessity to
extend the study beyond single random variable.
This chapter is devoted to a discussion on jointly related variables, their distribution
functions and other important characteristics. First we shall have a discussion on discrete
case.
(Dr. K. Gururajan, MCE, Hassan page 1)
-
8/10/2019 Stoc Proc KG
2/18
-
8/10/2019 Stoc Proc KG
3/18
Note: One caution with discrete random variables is that probabilities of events must be
calculated individually. From the preceding sections, it is clear that in the current
problem, there totally m n events. Thus, it is necessary to compute the probability of
each and every event. This can also be shown by means of table.
Y
X
y1 y2 y3 . . . . ny
x1 ,h x y1 1 ,h x y1 2 ,h x y1 3 . . . . , nh x y1x2 ,h x y2 1 ,h x y2 2 ,h x y2 3 . . . . , nh x y2
x3 ,h x y3 1 ,h x y3 2 ,h x y3 3 . . . . , nh x y3
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
mx ,mh x y1 ,mh x y2 ,mh x y3 . . . . ,m nh x y
Note: When the problem is based on discrete random variables, it is
necessary to compute the probability and each and every event
separately.
Illustrative examples:
A fair coin is tossed 3 times. Let X denote the random variable equals to 0 or 1
accordingly as a head or a tail occurs on the first toss, and let Y be the random
variable representing the total number of heads that occurs. Find the joint
distribution function of (X, Y).
Solution: S = {HHT, HHT, HTH, THH, THT, TTH, HTT, TTT}. Thus, |S| = 8.
Here, X takes the values 0 or 1 accordingly as a H appears on the I toss or a Tail appearson the I toss, while Y takes the values 0, 1, 2, 3 where these numbers represent the
number of heads appearing in the experiment. Observe that joint variables(X, Y) assume
eight values. Thus, there are (0, 0), (0, 1), (0, 2), (0, 3), (1, 0), (1, 1), (1, 2), (1, 3) totally 8 events. Thus, we need to find the probabilities of all these 8 events. First we
shall list the respective events corresponding to X and Y.
(Dr. K. Gururajan, MCE, Hassan page 3)
-
8/10/2019 Stoc Proc KG
4/18
[X = 0] = {HHH, HHT, HTH, HTT},
[X = 1] = {THH, TTH, THT, TTT}
[Y = 0] = {TTT},
[Y = 1] = {HTT, HTH, TTH],
[Y = 2] = {HHT, HTH, THH}
[Y = 3] = {HHH}
Therefore, [X = 0, Y = 0] = { }, a null event, hence [ , ] ( , )P X Y h0 0 0 0 0 .
Similarly, [ , ]X Y0 1 = {HTT}, so [ , ] ( , )P X Y h 10 1 0 1 8 . Note that
[ , ]X Y0 2 = {HHT, HTH}, thus [ , ] ( , )P X Y h 20 2 0 28
. Another example,
consider the event [X = 1, Y = 2] = {THH} implying that [ , ] ( , )P X Y h 11 2 1 28
.
Continuing like this, one can compute the probabilities of all the events. The results are
shown in the form of the following table;
Y
X
0 1 2 3
0 ( , )h 0 0 0 ( , )h 10 18
( , )h 20 28
( , )h 10 38
1 ( , )h 11 08
( , )h 21 18
( , )h 11 28
( , )h 1 3 0
Two cards are selected at random from a box which contains 5 cards numbered 1, 1,
2, 2, and 3. Let X denote the sum selected, and Y the maximum of two numbers
drawn. Determine the joint probability function of X and Y.
Solution: It is known that 2 cards can be selected from the available 5 in C (5, 2) = 10
ways. With cards being numbered as 1, 1, 2, 2, and 3, clearly, S = {(1, 1), (1, 2), (1, 2),
(1, 2), (1, 2), (1, 3), (1, 3), (2, 2), (2, 3), (2, 3)}. As X denote the sum of numbered
chosen, X takes the values 2, 3, 4, and 5. Y is defined as the maximum of the two
numbers, so Y takes the values 1, 2, 3. Therefore, (X, Y) assumes totally 12 values. It
can be shown as
(Dr. K. Gururajan, MCE, Hassan page 4)
-
8/10/2019 Stoc Proc KG
5/18
-
8/10/2019 Stoc Proc KG
6/18
Y
X
0 1 2 3
0 ( , )h 0 0 0 ( , )h 10 18
( , )h 20 28
( , )h 10 38
1 ( , )h 11 0 8 ( , )h 21 1 8 ( , )h 11 2 8 ( , )h 1 3 0
The distribution function of X is got by adding the probabilities row wise. Thus,
:
1 1:
2 2
i
i
x
f x
0 1
and the marginal distribution function of Y is to be obtained by the adding theprobabilities column wise. Therefore,
iy 0 1 2 3
jg y 183
83
81
8
The Co Variance of X and Y is given as Cov( , ) ( ) ( ) ( )X Y E XY E X E Y . The
regression coefficient between X and Y can be written asCov( , )
( , )x y
X Yx y
. Here,
and x y are standard deviation of X and Y series respectively.
ILLUSTRATIVE EXAMPLES:
Consider the joint distribution of X and Y
Y -4 2 7
1 1/8 1/8
5 2/8 1/8 1/8
Compute E(X), E(Y), E (XY), Covar (X, Y),y
,x and ( , ).x y
Solution: First, we obtain the distribution functions of X and Y. To get the distribution
of X, it is sufficient to add the probabilities row wise. Thus,
(Dr. K. Gururajan, MCE, Hassan page 6)
X
Y
-
8/10/2019 Stoc Proc KG
7/18
X 1 5
if x 1/2 1/2
( ) ( / ) ( / )E X 1 1 2 5 1 2 3
( / ) ( / )E X2 1 1 2 25 1 2 13
Var( ) ( )X E X E X22 4 and x 2 .
Similarly,
Y -4 2 7
jg y 3/8 3/8 2/8
( ) ( / ) ( / ) ( / ) .E Y 4 3 8 2 3 8 7 2 8 1 0
( / ) ( / ) ( / ) .E Y2 16 3 8 4 3 8 49 2 8 39 5
Var( ) ( )Y E Y E Y 22 = 38.5
.y 38 5 = 6.2048.
Consider ( ) ( , )x y
E XY x y h x y = (2)(1)(0.2) + (3)(2)(0.4) + (4)(2)(0.1) +
(4)(3)(0.2) + (5)(3)(0.2) = 8.8. With E(X) = 3, and E(Y) = 1, it follows that
Cov( , ) ( ) ( ) ( )X Y E XY E X E Y = 5.8 andCov( , ) .
( , ) .( . ) ( . )
X Yx y
x y
5 80 4673
2 0 6 2048
. As
Cov( , ) is not zero, we conclude that and are not Independent.X Y X Y
It may be noted that so far the study on probability theory was concentrated on either a
single random variable or a two random variables (correlated variables) defined on the
same sample space, S of some random experiment. It is interesting to see that in a
practical situation, we come across numerous occasions involving a family or a collection
(Dr. K. Gururajan, MCE, Hassan page 7)
-
8/10/2019 Stoc Proc KG
8/18
of random variables. This is especially true in problems of queuing theory and in
networking. As todays problems are mainly based on theory of queues, thus, there is a
strong necessity to study the same. Also, this has enormous amount of applications in
engineering and technology. With these in view, we shall consider a brief discussion on
stochastic processes. Among various types of stochastic processes available in literature,Markov processes have found significant applications in many fields of science and
engineering; therefore, current study will be restricted to this only.
A discussion on Independent Random Variables
Let X and Y be two discrete random variables. One says that X and Y are independent
whenever the joint distribution function is the product of the respective marginal
distribution functions. Equivalently, suppose that ( ) ( ) ( ),, g andi j i jf x y h x y denoterespectively, the distribution function of X, Y and that of X and Y,
and ( ) ( ) ( ), gi j i jh x y f x y= , then we say that X and Y are independent random
variables.
Note: If it is known that X and Y are independent random variables, then it is very easyto compute the joint distribution function. This may be obtained by just multiplying the
respective probabilities in the corresponding row and column of the table. For example,consider
X 0 1 2 3
( )if x 0.2 0.2 0.5 0.1
y 3 5 6
( )jg y 0.2 0.5 0.3
Then, joint distribution function of X and Y may be obtained as follows:
( ),i jh x y 3 5 6
0 0.04 0.10 0.06
1 0.04 0.10 0.06
2 0.10 0.25 0.153 0.02 0.05 0.03
(Dr. K. Gururajan, MCE, Hassan page 8)
-
8/10/2019 Stoc Proc KG
9/18
-
8/10/2019 Stoc Proc KG
10/18
not on the previous or past records. Mathematically, this can be explained as
( ) | , , . . .1 1 0n n n n o P X t x X t x X t x X t x
( ) |
n n
P X t x X t x
i.e. here, behavior of the stochastic process is such that probability distributions for its
future development depend only the present state and not on how the process arrived in
that state. Also, state space, I is discrete in nature.
Equivalently, in a Markov chain, (a Markov process where state space, I takes discrete
values), the past history is completely summarized in the current state, and, future isindependent of its past but depends only on present state.
First we shall discuss few basic concepts pertaining to Markov chains. A vector
, , . . . ,1 2 3
na a a a a is called a probability vector if all the components are nonnegative and their sum is 1.
A matrix, A such that each row is a probability vector, then A is called stochastic matrix.IfA and B are stochastic matrices, then AB is a stochastic matrix. In fact, any power of
A is a a stochastic matrix.
A stochastic matrix A is said to be regular if all the entries of A are positive for some
power mA of A where m is a positive integer.
Result 1: Let A be a regular stochastic matrix. Then A has a unique fixed probability
vector, t. i.e. One can find a unique fixed vector t such that t t A .
Result 2: The sequence , , , . . .2 3A A A of powers ofA approaches the matrix T whose
rows are each the fixed point t.
Result 3: Ifq is any vector, then , , , .2 3q A q A q A approaches the fixed point t.
Transition Matrix of a Markov Chain:
Consider a Markov chain a fine stochastic process consisting of a finite sequence of
trials, whose outcomes say, , , . . .1 2 3x x x satisfy the following properties:
(Dr. K. Gururajan, MCE, Hassanpage 10)
-
8/10/2019 Stoc Proc KG
11/18
Each outcome belongs to the state space, m{ , , . . . a }1 2 3I a a a . The outcome of any
trial depends, at most, on the outcome of the trial and not on any other previous
outcomes. If the outcome in thn trial isi
a then we say that the system is ini
a state at
the thn stage. Thus, with each pair of states ,i ja a , we associate a probability value
ijp which indicate probability of system reaching the state ia from the state ja in nsteps. These probabilities form a matrix called transition probability matrix or just
transition matrix. This may be written as
. . .
. . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . .
11 12 13 1
21 22 23 2
1 2 3
n
n
n n n n n
p p p p
p p p p
M
p p p p
Note: Here, thi row ofM represents the probabilities of that system will change from
ia to , , , . . . .1 2 3 na a a a Equivalently, | 1i j n n p P x j x i .
nstep transition probabilities:
The probability that a Markov chain will move from state I to state j in exactly n steps
is denoted by ( ) ( ) |ij
n
i j m n m p p n P x j x i
Evaluation of nstep transition probability matrix
If M is the transition matrix of a Markov chain, then the n step transition matrix may beobtained by taking nth power of M. Suppose that a system at time t = 0 is in the state
n, , . . . a1 2 3a a a a where the process begins, then corresponding probability vector
n
( ) ( ) ( ) ( ) (0), , . . . a1 2 3
0 0 0 0a a a a denotes the initial probability distribution. Similarly, the
step transition probability matrix may be given as n
( ) ( ) ( ) ( ) (n), , . . . a1 2 3
n n n n a a a a . Now,
the (marginal) probability mass function of the random variable may be obtainedfrom the nth step transition probabilities and the initial distribution function as follows:
( ) ( ) ( ) ( )0 0n n na a M a M . Thus, the probability distribution of a homogeneous Markov
chain is completely determined from the one step transition probability matrix M and theinitial probability distribution (0 )a .
Note: The same may be explained as
(Dr. K. Gururajan, MCE, Hassanpage 11)
-
8/10/2019 Stoc Proc KG
12/18
( ) ( )1 0a a M
( ) ( ) ( )2 1 0 2a a M a M
( ) ( ) ( )3 2 0 3a a M a M
Some illustrative examples on Markov chains:
1. Determine which of the following are stochastic matrices?
(i)
1 1 1
3 3 3
1 1 02 2
(ii)
3 1
4 4
2 23 3
(iii)
3 1 -
4 4
2 13 3
Solution: (i) is a stochastic matrix as all the entries are nonnegative and the sum of allvalues in both the rows equals 1. (ii) This is not a stochastic matrix, since the second row
is such that sum exceeding 1 (iii) this is again not a stochastic matrix, because one of the
entry is negative.
2. Determine which of the following are regular stochastic matrices?
(i)
1 1 1
2 4 4
0 1 0
1 1 0
2 2
A
(ii)
1 1 1
2 4 4
0 0 1
0 1 0
B
Solution: A is not a regular stochastic matrix, as it has a 1 on the main diagonal. B is a
regular stochastic matrix, Compute 3B , you would notice that all the entries are positive.
3. Find the unique fixed probability vector of the following stochastic matrices.
(i)23
0
13
1
A
(ii)
3 1
4 4
1 0
2
1 0
0
1
2
0
(iii)
1 0
1 1
2 3
2 1
3 3
0
1
6
0
A
(Dr. K. Gururajan, MCE, Hassanpage 12)
-
8/10/2019 Stoc Proc KG
13/18
Solution: Here, to find a unique fixed probability vector ( , ) such thatt x y t t A .
As t is a probability vector, one can take ( , )1t x x . Consider the matrix equation,
2
( , ) ( , ) 3
0
1
1 1 3
1
x x x x
; A multiplication of RHS matrices, yields,
( )1
13
x x x and ( )2
13
x x . Thus, one can find that .0 6x , hence the
required unique fixed probability vector ( . , . )0 6 0 4t .
(ii) We shall set up ( , , )t x y z as the unique fixed probability vector. Again this is a
probability vector, therefore, ( , , )1t x y x y . Now, consider the matrix equation,
3 1
4 4
1( , , ) ( , , ) 02
1 0
0
11 12
0
x y x y x y x y
Multiplying the matrices on the right hand side gives,
1
2x y
( )3 1
1
4 2
y x y x y
Solving these two equations, one obtains the unique fixed probability vector.
(iii) Here, the unique fixed probability vector may be obtained as ( . , . , . )0 1 0 6 0 3t .
5. A boys studying habits are as follows. If he studies one night, he is 70 percent
sure not to study the next night. On the other hand, if he does not study one night,
he is only 60 percent sure not to study the next night as well. In the long run, how
often does the boy studies?
Solution: First we construct transition probability matrix. The problem here is about
whether the boy studies or not on consecutive nights. As his study is given to bedependent on study in one night or not, therefore, it may be modeled in terms of Markov
chain. Thus, first we construct the required transition probability matrix. The following
table indicates the pattern of the study of the boy:
(Dr. K. Gururajan, MCE, Hassanpage 13)
-
8/10/2019 Stoc Proc KG
14/18
I night
II night
Transition probability Boy Studying Boy not studying
Boy Studying 0.3 0.7
Boy not studying 0.4 0.6
Therefore, it is clear that required stochastic matrix is0.3 0.7
0.4 0.6M
. To solve the
given problem, it is sufficient to compute a unique fixed probability vector, ( , )1t x x
such that t t M . Using the above, one obtains . . ( )0 3 0 4 1x x x . From this
equation, we can calculate4
11x . Thus, a chance of boy studying in the long run is
36.7%.
6. A persons playing chess game or tennis is as follows: If he plays chess game one
week then he switches to playing tennis the next week with a probability 0.2. On the
other hand, if he plays tennis one week, then there is a probability of 0.7 that he will
play tennis only in the next week as well. In the long run, how often does he play
chess game?
Solution: Like in the previous case, this one too is based on Markov process. Here,
parameters are about a player playing either chess or tennis game. Also, playing the
game in the next week is treated to be dependent on which game player plays in theprevious week. As usual, first we obtain transition probability matrix.
I Week
II Week
Transition
probabilities
Player Playing Chess Player Playing Tennis
Player Playing
Chess 0.8 0.2
Player Playing
Tennis 0.3 0.7
(Dr. K. Gururajan, MCE, Hassan, page number 14)
-
8/10/2019 Stoc Proc KG
15/18
Clearly, the transition probability matrix is0.8 0.2
0.3 0.7M
. Here, too, the problem is
to find a unique fixed probability vector, ( , )1t x x such that t t M . Consider the
matrix equation;0.8 0.2
- ) ( - )
0.3 0.7
1 1x x x x
. A simple multiplication work
results in . . ( )0 8 0 7 1x x x from which we get .0 6x . Hence, we conclude that in
the long run, person playing chess game is about 60%.
7. A sales man S sells in only three cities, A, B, and C. Suppose that S never sells in
the same city on successive days. If S sells in city A, then the next day S sells in the
city B. However, if S sells in either B or C, then the next day S is twice likely to sell
in city A as in the other city. Find out how often, in the long run, S sells in each city.
Solution: First we shall obtain transition probability matrix which will have 3 rows and
3 columns. As a salesman does not sell in a particular city on consecutive days, clearly,
main diagonal is zero. It is given that if S sells in city A on first day, then next day hewill sell in city B, therefore, first row is 0, 1, 0. next, one can consider two cases: (i) S
selling city B or (ii) S selling in city C. It is also given that if S sells in city B with
probability p, then his chances of selling in city A, next day is 2p. There is no way, hewill sell in city C. Thus, we have p + 2p + 0 = 1 implying that p = 1/3. Therefore middle
row of the matrix is 2/3, 1/3, 0. Similarly, if S sells in city C with probability q, then his
chances of selling in city A, next day is 2q. Again, a chance of selling city B is 0. Thus,last row is 2/3, 0 , 1/3. The probability transition matrix is
I day
Next Day
A B C
A 0 1 0
B 23
0 13
C 23
13
0
(Dr. K. Gururajan, MCE, Hassan, page number 15 , 2405 2-008)
-
8/10/2019 Stoc Proc KG
16/18
The matrix of the Markov process is
0 1 0
2 1 03 3
2 1 03 3
M
. As in the previous two
problems, we need to find a unique fixed probability vector, ( , , )t x y z such that
0 1 0
2 1 1 - - 1 - - 03 3
2 1 03 3
x y x y x y x y
Or
2 2 (1 )3 3
x y x y and 1 (1 )3
y x x y . Solving these two equations,
one obtains 0.4, 0.45 and 0.15x y z . Hence, chances of a sales man selling in
each of the cities is 40%, 45% and 15% respectively.
8. Marys gambling luck follows a pattern. If she wins a game, the probability of
winning the next game is 0.6. However, if she loses a game, the probability of losing
the next game is 0.7. There is an even chance that she wins the first game. (a) Find
the transition matrix M of the Markov process. (b) Find the probability that she
wins the second game? (c) Find the probability that she wins the third game? Find
out how often, in the long run, she wins?
10. Each year Rohith trades his car for a new car. If he has a Maruti, he trades it in
for a Santro. If he has a Santro car, then he trades it in for a Ford. However, is he
has a Ford car; he is just as likely to trade it in for a new Ford as to trade it in for a
Maruti or for a Santro. In 1995, he bought his first car which was a Ford.
(a) Find the probability that he has bought (i) a 1997 Buick, (ii) a 1998 Plymouth,(iii) a 1998 Ford?
(b) Find out how often, in the long run, he will have a Ford?
Definition of various states:
Transient state: A state i is said to be transient (or nonrecurrent) if and only if there is
a positive probability that process will not return to this state. For example, if we modela program as a Markov chain all but the final state will be transient. Otherwise program
ends in an infinite loop.
Recurrent state: A state i is said to be recurrent if and only if process returns to this
state with probability one.
Periodic state: A recurrent state which returns to the state i at regular periods of time.
Absorbing state: A state i is said to be absorbing if and only if 1ii
p . Here, once a
Markov chain enters a state, and it remains there itself.
(Dr. K. Gururajan, MCE, Hassan, page number 16, 2405 2-008)
-
8/10/2019 Stoc Proc KG
17/18
PLEDGE
I commit to excel, in all I do.I will apply myself actively to making a difference at
Malnad College of Engineering, Hassan.
I will use every opportunity presented to me by my
superiors from this moment to make that difference.
For myself, for colleagues, and for my students
(Dr. K. Gururajan)
-
8/10/2019 Stoc Proc KG
18/18