Professional Documents
Culture Documents
that for distinct values of n (=1,2,3…), one gets distinct random variable X n ;
State space: The set of possible values of a single random variable X n of a stochastic
of possible values of X n is the finite set of non negative integers 0,1,…n Hence the
assuming values in [0, ). Here, the set of values of X n is the interval [0, ), and so
All the four types may be represented by { X (t ), t T } . In case of discrete time, the
parameter generally used is n, i.e,. the family is represented by { X (n), n 0,1,2....} . In
case of continuous time both the symbols { X t , t T } and { X (t ), t T } ( where T is
finite or infinite interval) are used. The parameter t is usually interpreted as time,
through it may represent such characters as distance, length, thickness and soon.
Markov process: If { X (t ), t T } is a stochastic process such that, for,
t1 t 2 ... t n t
j, k , j1 , j n1 N
(i.e. X n j ) , the process is said to be at state j at the nth trail. To a pair of states (j, k)
at the two successive trails ( nth and (n+1)th trails) there is an associated conditional
probability P jk .
probability of transition from state j at the nthtrial to the state k at the (n+1)th trail.
Homogeneous Markov chain: If the transition probability P jk is independent of n,
k) at two successive trails (say nth and (n+1)th trails); the transition is one step
transition probability.
If we are concerned with the pair of states (j, k) at two non-successive trails,
say, j at the nth trail and k at the (n+m)th trail, the corresponding probability is then
PrX n m k / X n j.
( m)
called m-step transition probability and is denoted by Pjk
This is called the transition probability Matrix of the Markov chain. P is a stochastic
matrix.
Thus a transition matrix is a square matrix with non-negative elements and unit-row
sums.
Example 1: Consider a communication system which transmits the two digits 0 and 1
through several stages. Let X n , n 1 be the digit leaving the nth stage of system and
X0 be the digit entering the first stage. At each stage there is a constant probability q
that the digit which enters will be transmitted unchanged, and probability p otherwise,
p+q=1.
Here X n , n 0 is a homogeneous two-state Markov chain with single step transition
matrix
0 1
0 q p
P
1 p q
Example 2. A particle performs a random walk with absorbing barriers say, as 0 and
4. Whenever it is at any position r (0 < r < 4), it moves to r+1 with probability p or to
(r-1) with probability q, p+q=1. But as soon as it reaches 0 or 4 it remains there itself.
Let Xn be the position of the particle after n moves. The different states of Xn are the
different positions of the particle {Xn} is a Markov chain whose unit-step probabilities
are given by
PrX n r 1 / X n 1 r p
PrX n r 1 / X n 1 r q, 0 r 4
and
PrX n 0 / X n 1 0 1
PrX n 4 / X n 1 4 1
The transition matrix is given by
States of X n
0 1 2 3 4
0 1 0 0 0 0
1 q 0 p 0 0
States of X n 1 2 0 q 0 p 0
3 0 0 q 0 p
4 0 0 0 0 1
Eg 3: General case of random walk: In the above example, as soon as the particle
reaches 0 it remains there with probability ‘a’ and is reflected to state 1 with
probability 1-a(0 < a <1); if it reaches state 4 it remains there with probability b and
reflected to state 3 with probability 1-b (o < b<1) then {Xn} is a Markov chain with
state space {0,1,2,3,4}.
The transition matrix is given by
0 1 2 3 4
0 a 1 a 0 0 0
1 q 0 p 0 0
2 0 q 0 p 0
3 0 0 q 0 p
4 0 0 0 1 b b
If a=1, then ‘0’ is an absorbing barrier and if a=0, then ‘0’ is a reflecting barrier
similar is a case with state 4.
Finite Markov chain: A Markov chain X n , n 0 with k states, where k is finite, is
said to be a finite Markov chain. The transition matrix P in this case is a square matrix
with k rows and k columns.
The number of states could however infinite. Where the possible values of X n form a
denumerable set, then the Markov chain is said to be denumerable infinite or
denumerable and the chain is said to have a countable state space.
Probability distribution:
The probability distribution of X 0 , X 1 , X 2 , , X n can be computed in terms of the
PrX 0 a, X 1 b, , X n 2 i, X n 1 j , X n k
PrX n k / X n 1 j , X n 2 i, , X 1 b, X 0 a
PrX n k / X n 1 j prX n 1 j , X n 2 i, , X 0 a
PrX n k / X n 1 jprX n 1 j / X n 2 i PrX 1 b / X 0 aPrX 0 a
p jk p ij p ab PrX 0 a.
Eg: Let X n , n 0 be a Markov chain with three states 0,1,2 with transition matrix
0 1 2
0 3 1
4 0
4
1
1 1 1
2 4 2 4
3 1
0
4 4
and the initial distribution PrX 0 i , i 0,1, 2.
1
3
We have PrX 1 1 / X 0 2
3
4
PrX 2 2 / X 1 1
1
4
PrX 2 2, X 1 1 / X 0 2
PrX 2 2 / X 1 1PrX 1 1 / X 0 2 .
1 3 3
4 4 16
PrX 3 1 / X 2 2Pr X 2 2, X 1 1, X 0 2 .
3 1 3
4 16 64
Note:The transition probabilities with the initial distribution completely
specifies a
Markov chain.
Higher order transition probabilities:
PrX n 2 k / X n j PrX n 2 k , X n 1 r / X n j
( 2)
p jk
r
p jr p rk
r
p rk p Jr
( m n) (n) ( m) (n) ( m)
p jk p jr p rk
r
Let p ( p jk ) denote the transition matrix of the unit-step transition and P ( m) Pjk ( m)
denote the transition matrix of the m-step transitions. For m=2, we have the matrix
p ( 2) p. p p 2 similarly p ( m n) p ( m) p p. p ( m)
Classification of chains: The Markov chains are of two types (i) ergodic (ii) regular
An ergodic Markov chain has the property that it is possible to pass from one state to
another in a finite number of steps, regardless of present state.
A special type of ergodic Markov chain is the regular Markov chain.
A regular Markov chain is defined as a chain having a transition matrix P such that for
some power of P it has only non-zero positive probability values.
Note: Thus all regular chains must be ergodic chains.
Example: Consider a communication system which transmits the digits 0 and 1
through several stages. At each stage the probability that the same digit will be
received by the next stage, as transmitted, is 0.75. What is the probability that a 0 that
is entered at the first stage is received as a 0 by the 5th stage?
0.75 0.25
The state transition matrix P is given by
P 0.25 0.75
0.625 0.375
Hence P 0.375 0.625
2
0.53125 0.46875
And
P 4
P 2 2
P 0.46875 0.53125
Therefore the probability that a zero will be transmitted through four stages as a zero is
P004=0.53125.
Session-2
Problems to be discussed by the faculty:
1) Which of the following matrices are T.P.M / stochastic matrices
1 2 4
1 0 0
a. [ ] b. [31 3 3
1]
0 1 0 1
2 2
15 1
1 0
𝑐. [ ] d. [16
2
16
4]
0 1
3 3
1 0 0 2
e. [ ] f. [ 1 1 ]
−1 0 4 4
1
0 2
g. [ 1 1]
3 4
2 3 4 5
0 0 1 1
0
1 1
c) [ 2 0 2
] d) [31 ]
1
0 1 0 3
1 1
1
2 4
1 0
e) [ ] f) [0 1]
1
0 1 2
0 0 1
6) Using the diagram write the answers to the following
a. Find P(X4=3|X3=2)
b. Find P(X3=1|X2=1)
c. If we know P(X0=1)=13 find P(X0=1,X1=2).
d. If we know P(X0=1)=13, find P(X0=1,X1=2,X2=3)
7) Which of the stochastic matrices are ergotic
0 1 3 −1 1 0
1 0
𝑎) ( ) 𝑏) (2 0 1) 𝑐) ( 1 0 1)
0 1
1 −2 0 1 −1 1
At present, it is estimated that 60% of the people buy brand A and 40% buy brand
B. Determine the market shares of brand A and brand B in the steady state.
5. Consider a certain community in well-defined area with three types of grocery
stores; for simplicity we shall call them I, II and III. Within this community (we
assume that the population is fixed) there always exists a shift of customer from
one grocery store to another. A study was made on January 1 and it was found that
¼ shopped at store I, 1/3 at store II and 5/12 at store III. Each month store I retains
90% of its customers and loses 10% of them to store II. store II retains 90% of its
customers and loses 5% each to store I store III. Store III retains 40% of its
customers and loses 50% of them to store I and 10% to store II.
I) What proportion of customers will each store retain by February 1;
March 1?
II) Assuming the same pattern continues, what will be the long run
distribution of customers among the three stores?
6. Three children (denoted by 1,2,3) arranged in a circle play a game of throwing a
ball to one another. At each stage the child having the ball is equally likely to
throw it into any one of the other two children. Suppose that X 0 denote the child
who had the ball initially and X n (n 1) denotes the child who had the ball after n
throws. Show that { X n , n 1} forms a Markov chain. Find p. Calculate
PrX 2 1 / X 0 1. , PrX 2 2 / X 0 3 , PrX 2 3 / X 0 2 and also the
probability that the child who had originally the ball will have it after 2 throws.
Find p if the number of children is m ( 3).
Session 4
1. A market research team has conducting a survey of customer buying habits with
respect to three brands of talcum power in an area. It estimate, at present, 20% of the
customers buy brand A, 50% of the customers buy brand B and 30% of the customers
buy brand C. In addition the team has analyzed its survey and has determined the
following brand switching matrix.
Brand Next Bought A B C
What will be the expected distribution of consumers two time periods later and also at
equilibrium.
2. A research analyzing brand switching between different airlines, operating on the
Delhi-Mumbai route by frequent fliers. On the basis of the data collected by her, the
researcher has developed the following transition probability matrix.
To airline
AA 0.9 .03 0.07
From airline BB 0.15 0.80 0.05
CC 0.20 0.30 0.50
It is found that currently the airlines AA, BB and CC have 20%, 50% and 30% of the
market respectively.
i) Obtain the market share for each airline in two moths time, and
ii) Calculate the long run market share for each time.
3. Honey lnc had 35% of the local market for its cosmetics, while the two other
manufactures of cosmetics Lace Lnc and shalon lnc have 40% and 25% shares
respectively in the local market as on Ist April of this year. A study by a market
research firm has disclosed the following:
Honey Inc. retains 86% of its customers, while it gains 4% and 6% of the customers
from its two competitors, Lace and Shalon respectively. Lace. Inc. retains 90% of its
customers, and gains 8% and 9% of customers respectively., from Honey and Shalon.
Shalon retains 85% of its customers and gain 6% and 6% of customers from Lace and
Honey, respectively.
Number of customers/day 8 10 12 14
Probability 0.35 0.30 0.25 0.10
Customer order 1,2,3 or 4 dozens bagels according to the following distribution
Number of dozens order/customer 1 2 3 4
Probability 0.4 0.3 0.2 0.1
Bagel sells for $8.40 per dozen. They cost $5.80 per dozen to make. All bagels not
sold at the end of the day are sold at half-price to a local grocery store. Based on 5
days of simulation, how many (to the nearest 5 dozens) bagels should be baked each
day?
2) There is only one telephone in a public booth of a railway station. The following table
indicate the distribution of callers arrival time and duration of the calls
Call duration(minutes 2 3 4 5
Probability 0.15 0.6 0.15 0.1
Simulate for the current system. It is proposed to add another telephone to the booth.
Justify the proposal based on the waiting time of callers.
Session-6
3) The newsstand buys the papers for 33 cents each and sells them for 50 cents each.
Newspapers not sold at the end of the day are sold at scrap for 5 cents each.
Newspapers can be purchased in bundles of 10. Thus the newsstand can buy 50, 60,
and so on. There are three types of news days: ‘good’; ‘fair’; and ‘poor’; they have
the probabilities 0.35, 0.45 and 0.20 respectively. The distribution of newspapers
demanded on each of these days is given in the following table
Demand probability distribution
Demand Good Fair poor
If the decision is to purchase 50 newspapers per day simulate the demand for 8 days
and determine the total profit.
4) A milling machine has three different bearings that fail in service. The distribution
of the life of each bearing is identical as given in the following table. When a bearing
fails, the mill stops, a repair person is called, and a new bearing is installed.
Bearing life(hours) 100 110 120 130 140
0 0 0 0 0
Probability 0.2 0.2 0.2 0.1 0.1
0 3 8 9 0
The probability distribution of the delay time of the repair person is given in
following table.
Delay(minutes) 5 10 15
Probability 0.6 0.3 0.1
Downtime for the mill is estimated as $10 per minute the on-site cost of the repair
person is $30 per hour. It takes 20 minutes to change one bearing, 30minutes to
change 2 bearings and 40 minutes to change 3 bearings. A proposal has been made to
replace all bearings when ever a bearing fails. The cost of each bearing is $35.
Simulate 5 bearings changes to evaluate the proposal using the total cost per 5000
bearing hours.
5) Students arrive at the University library counter with inter arrival times
distributed as
Time between arrivals(minutes) 5 6 7 8
Sales of cars 37 38 39 40 41 42
Probability 0.10 0.15 0.2 0.3 0.1 0.0
0 5 5 5
The production cost and sale price of each are Rs. 4 lacks and Rs 5 lakh
respectively. Any unsold car is to be disposed off at a loss of Rs. 2 lakh per car. There
is a pently of Rs 1 lakh per car, if the demand is not met. Using the following random
numbers, estimate total profit/loss for the company for the next ten days:
9, 98, 64, 98, 94, 01, 78, 10, 15, 19.
If the company decides to produce 39 cars per day, what will be its impact on
profitability?
7. The director of finance for a farm co-operative is concerned about the yields per acre
she can expect from this year’s corn crop. The probability distribution of the yields
for the current weather conditions is given below:
(ii) Simulate the price she might expect to observe over the next 10 years using
the following random numbers: 82, 95, 18, 96, 20, 24, 56, 11, 52, 3.
(iii) Assume that prices are independent of yields combine these two into the
revenue per acre and also find out the average revenue per acre she might expect every
year.
8. A firm has a single channel service station with the following arrival and
service time probability distributions:
Inter-arrival PROBABILITY Service time PROBABILITY
time (minutes) (Minutes)
10 0.10 5 0.08
15 0.25 10 0.15
20 0.30 15 0.18
25 0.25 20 0.24
30 0.10 25 0.22
30 0.14
The customers arrival at the service station is a random phenomenon and the time
between the arrivals varies from 10 minutes to 30 minutes. The service time varies from
5 minutes to 30 minutes. The queuing process begins at 10AM and proceeds for nearly
8 hours An arrival goes to the service facility immediately, if it is free. Otherwise, it
will wait in a queue. The queue discipline is first-come first served. If the attendant’s
wages are Rs. 10 per hour and the customer’s waiting time costs Rs. 15 per hour, then
would it be economical proposition to engage a second attendant? Answer using Monte
- carlo simulation technique.