Markov Analysis

Example: In a certain market, only two brands of cold drinks; Pepsi and Coke. Given
last purchased Pepsi, there is 80% chance that he would buy the same brand in the n
while if his current purchase is Coke, there is 90% chance that he would buy Coke
purchase. Create a transition probability matrix and determine

(i) The probability that if a customer is currently a Pepsi purchaser, he will p

two purchases from now

(ii) The probability that if a customer is currently a Coke purchaser, he will pu

three purchases from now

Transition Probability Matrix

Solution:
Next
Pepsi Coke
Current Pepsi 0.8 0.2 P
Coke 0.1 0.9

Solution: Initially customer is either Pepsi or coke which can be represented by

Pepsi Coke
Pepsi 0.8 0.2
Coke 0.1 0.9

(i) Currently he is Pepsi purchasPepsi Coke

Pepsi 1 0

S(0) = 1 0

After a year
S[1] S[0] x P

[1 0] 0.8 0.2
 0.1 0.9

1x0.8+0x0.1 1x0.2 +0x 0.9

Pepsi Coke
S[1] = 0.8 0.2

if currently customer is pepesi purchaser, after a year, there is 80% chance that he will
switch to Coke.
After two year,

S[2] = S[1] x P
0.8 0.2
0.8 0.2 0.1 0.9

0.8x0.8+0.2x0.1 0.8x0.2+0.2x0.9
Pepsi Coke
0.66 0.34

if currently customer is pepesi purchaser, after two year, there is 66% chance that he w
switch to Coke.

3 purchases from now if currently coke purchaser

21.9% to Pepsi and 78% to Coke
Pepsi and Coke. Given that a man
e same brand in the next purchase,
t he would buy Coke in the next
ix and determine

si purchaser, he will purchase Coke

now
e purchaser, he will purchase Pepsi
now

e represented by
% chance that he will retained in the same brand and 20 % chance that he will

66% chance that he will retained in the same brand and 34 % chance that he will
Problem 2. In a survey, it was found that the mobility of the population
of a state to a village, town, and city is in the following percentages
What will be the proportion in village, town and city after two years
given that the present population has proportion of 0.7, 0.2, 0.1 in the
village, town and city respectively?
Given :
Village Town
Village 0.5 0.3
Town 0.1 0.7
City 0.1 0.4

Solution : Current proportion is given as

V T
0.7 0.2

After a year the proportion of village, town and city is

S(1) = 0.7 0.2

S(1) = 0.38 0.39

After two year the proportion of village, town and city is

S(2) = S(1) x P

0.38 0.39

Village Town
S(2) = 0.252 0.479
After two year the population proportion will be 25.2% in Village , 47.9% i
0.1 in the

City
0.2
0.2 = P
0.5

C
0.1

0.5 0.3 0.2

0.1 x 0.1 0.7 0.2
0.1 0.4 0.5

0.23

0.5 0.3 0.2

0.23 x 0.1 0.7 0.2
0.1 0.4 0.5
City
0.269
Village , 47.9% in Town and 26.9% in City.
Long Run or Steady State

Problem 3.      New razor blades were introduced in the market by three companies at
were introduced, each company had an equal share of the market, but during the first ye
place:
(i)     Company A retained 90% of its customers, lost 3%to B and 7% to
(ii)  Company B retained 70% of its customers, lost 10% to A and 20%
(iii)   Company C retained 80% of its customers, lost 10% to A and 10%
Assuming that no changes in buying habits of consumers occur, determ
(a)   What are the market shares of the three companies at the end of first year?
(b) What are the long-run/ Steady state market share of the three companies

Solution Transition Probability Matrix

A B C
P= A 0.9 0.03 0.07
B 0.1 0.7 0.2
C 0.1 0.1 0.8

The current market share of three companies is

S[0] = [ 0.333 0.333 0.333]

At the end of Ist year, the market share of company A, B and C can be obtained as
S[1] = S[0]x P
Answer :

(b) Let the steady state probabilities are x, y and z for company A, B and C respectivel

x + y + z = 1…….(1)

We know that if steady state probabilites are multiplied with transition probability m
it will remain same (does not change), therefore
0.9 0.03
[x y z] = x y z x 0.1 0.7
0.1 0.1

[x y z] = [0.9x +0.1y+0.1z 0.03x+0.7y+0.1z 0.07x+0.2y+0.8z

We know that if two matrices are equal, their corresponding element will also be equ

x= 0.9x +0.1y +0.1z ====> x - 0.9x -0.1y - 0.1 z = 0 ===> 0.1 X- 0.1Y -0.1Z = 0…
y = 0.03x + 0.7 y + 0.1z====> -0.03X + 0.3Y -0.1 Z = 0……..(3)
z= 0.07x +0.2y + 0.8 z =====> -0.07 X -0.2Y + 0.2Z = 0 ……..(4)

now we can equate equation 2 & 3 (since its RHS are equal)

0.1 X- 0.1Y -0.1Z = -0.03X + 0.3Y -0.1 Z

0.13X -0.4 Y +0.Z = 0 …………(5)

from eqn 3 & 4
0.04 X +0.5Y -0.3 Z = 0……..(6)

From ratio property, we can find with the help of equation 5 & 6

Now,
 𝑿/((−𝟎.𝟒 𝑿−𝟎.𝟑)−(𝟎𝑿𝟎.𝟓)) = (−𝒀)/((𝟎.𝟏𝟑𝑿
𝒁/(𝟎.𝟏𝟑𝑿𝟎.𝟓 −(−𝟎.𝟒𝑿𝟎.𝟎𝟒) )
𝑥/0.12= (−𝑦)/(−0.039)=𝑧/0.081=
(𝑥+𝑦+𝑧)/(012+0.039+0.081)

Equate each ratio with the last ratio

X/0.12 = 1/0.24 ===> X = 0.12/0.24 ===> X = 0.5

Y/0.039 = 1/ 0.24 ===> Y = 0.1625

Z /0.081 = 1/ 0.24 ===> Z = 0.3375

In long run, the market share of the company A will 50 %, B will be 16.25% and for co
y three companies at the same time. When they
ut during the first year, following changes took

st 3%to B and 7% to C.
t 10% to A and 20% to C
st 10% to A and 10% to B.
umers occur, determine:
nd of first year?
ee companies

be obtained as
, B and C respectively.

nsition probability matrix

0.07 order of matrix = no. of rows of Ist matrix x no.of col of 2nd m
0.2
0.8

0.07x+0.2y+0.8z]

ment will also be equal, therefore

.1 X- 0.1Y -0.1Z = 0………(2)

0.13X -0.4 Y +0.Z = 0 …………(5)

0.04 X +0.5Y -0.3 Z = 0……..(6)

−𝒀)/((𝟎.𝟏𝟑𝑿(−𝟎.𝟑))−𝟎𝑿𝟎.𝟎𝟒)) =

e 16.25% and for company C, it will be 33.75 %.

o.of col of 2nd matrix
Problem 4. A college canteen cooks three kinds of sweets S1, S2 and S3. It is known th
never cooks the same sweet on successive weeks. If college canteen cooks S1 then ne
However, if college canteen cooks S2 or S3 then next week college canteen is twice a
than the other sweet. Obtain transition probabilty matrix and determine how often c
sweets in long run/ Steady state.

Solution: Transition Probability Matrix

Next
S1 S2 S3
S1 0 0 1
Current S2 0.67 0 0.33
S3 0.67 0.33 0

Long run probabilities, let it is x, y and z respectively for sweet S1, S2 and S3.

S= x y z
In steady state

[X Y Z ] =[ X Y Z]
𝑥/(−1.667.(−1.333)− (−0.334).(1.333))= (−𝑦)/(1(−1.333)−(−0

𝑥/(2.2221−(−0.4452))= 𝑦/(−0.334+1.333))=𝑧/(1.333+1.667)

𝑥/2.667= 𝑦/0.999=𝑧/3= 1/6.666

x= 0.4 , y= 0.15 , Z=
2 and S3. It is known that college canteen
nteen cooks S1 then next week, they cook S3
ege canteen is twice as likely to cook S1
determine how often college canteen would cook each of the

S1, S2 and S3.

𝑦)/(1(−1.333)−(−0.334)(1))=𝑧/((1). (1.333)−(1. (−1.667))

)=𝑧/(1.333+1.667))

0.45
Problem 5. A salesman's territory consist of three cities, A, B and C. He never sells in t
on successive days. If he sells in city A then next day he sell in city B. However, if he s
then the next day he is thrice as likely to sell in city A as in the either city. In the long
in each of the cities?

A B C
A 0 1 0
B 0.75 0 0.25
C 0.75 0
He never sells in the same city
However, if he sells in either B or C
city. In the long run, how often does he sell
Example: In a certain market, only two brands of cold drinks; Pepsi and Coke. Given
last purchased Pepsi, there is 80% chance that he would buy the same brand in th
purchase, while if his current purchase is Coke, there is 90% chance that he would a
Coke in the next purchase. Create a transition probability matrix

Solution: Transition Probability Matrix

Next
Pepsi Coke
Current Pepsi 0.8 0.2 P
Coke 0.1 0.9
psi and Coke. Given that a man
the same brand in the next
ance that he would again buy
bability matrix
 A B

1 2 5 6
7 8

For matrix multiplication:

the number of column in Ist matrix equal to the number of rows in the 2nd

AxB= 1x5+2x7 1x6+2x8 19 22

3x5+4x7 3x6+4x8 43 50

Order of matrix = No. of Rows in Ist matrix xNo. Of col in 2nd matrix
of rows in the 2nd matrix