!

!
Module 1
!
Lecture 3 : Random Vibrations & Failure Analysis
Introduction to Probability-I
!
!
!
Sayan Gupta
Department of Applied Mechanics
Indian Institute of Technology Madras
Introduction
Let us perform an experiment of tossing a coin.
!
The possible outcomes of are either heads (H) or tails (T).
!
If we toss the coin 5 times, the possible sequence of outcomes could
be { H H T H T }.

Terminologies:
Experiment: The tossing of the fair coin.
!
Sample Event: Every thinkable outcome of the experiment is completely
described by one, and only one, sample point. Here. the sample points are
either H or T.
!
Sample space: The sample space consists of all the possible sample points
of the experiment and is denoted by ⌦ . Here, ⌦ = { H, T }.
!
Outcome or Event:This is an aggregate of the sample points, eg. {H, H, T, H, T }

2
Experiment 1: Tossing a coin twice

Sample space is ⌦ = {HH, HT, T H, T T }

!
The sample space is finite and the number of elements in the sample space
is given by n(⌦) = 4
!
!
A sample point is HH, HT , T H or T T .
!
!
Here, the first element is the outcome of the first toss while the second
element is the outcome of the second toss.
!

3
Experiment 2: Tossing a dice.

Sample space ⌦ = {1, 2, 3, 4, 5, 6}.

!
The sample size is finite with countably finite elements.

Experiment 3: Tossing a coin repeatedly and counting the number of tosses

required until the first H appears.

Sample space = {1, 2, 3, 4, ...}

!
There are infinite number of outcomes. Hence, the sample space is
countably infinite.

In Experiments 1, 2 and 3, the sample space consists of discrete sample

points and hence are countable.

4
Experiment 4: Measuring the lifetime of a machine component.

Sample space
!
!
⌦ = {t; 0  t  1}
!
The sample space consists of continuos sample points between time
t = 0 to t = 1.
!
Since the sample space is continuous, the number of sample points
are uncountable.
!
Also, since the sample space is unbounded on one side, the sample
size is uncountably infinite.

An example of a sample space which is uncountably finite is the

experiment defined as the x-coordinate of a rain drop falling on a
square tile of length a .
!
The sample space is ⌦ = {x; 0  x  a}
5
Let us define an experiment of throwing a dice.
!
The sample space is ⌦ = {1, 2, 3, 4, 5, 6}
!
Let us define an event A = odd number appears
!
The sample space is
A = {1, 3, 5}
!
!
Another event B = even number appears
!
The sample space is ⌦B = {2, 4, 6}

⌦A ⇢ ⌦ implies that ⌦A is a subset of ⌦.

⌦B ⇢ ⌦ implies that ⌦B is a subset of ⌦.

⌦A and ⌦B are mutually exclusive.

Thus, ⌦ = {⌦A , ⌦B }
6
Example:
!
Let us consider another example.
Let
A = { no. of tosses until first head appaers }
⌦A = {1, 2, 3, ...}

B = { no. of tosses till first head is odd }

⌦B = {1, 3, 5, ...}

C = { no. of tosses required is less than 5 }

⌦C = {1, 2, 3, 4}
It is obvious that ⌦B ⇢ ⌦A and ⌦C ⇢ ⌦A .
!
However, the events B and C are not mutually exclusive.

The events and their relationship among each other cab be best represented
pictorially through what are called Venn Diagrams.
These are discussed in the next slide. 7
Venn Diagrams

•The square box represents the universal

set ⌦A, i.e, the set of all possible
outcomes of the experiment.
!
• The circle and ellipse represents the sub
sets ⌦B and ⌦C.
!
• Since ⌦B and ⌦C are not mutually
exclusive, they have an intersection
consisting of the numbers 1 and 3 which
belong to both ⌦B and ⌦C .

8
Venn Diagrams
Some common representations of Venn Diagrams are as follows:

A[B Ac

The darker shaded The shaded region

region is A \ B represents A \ B c
9
Venn Diagrams
A : occurrence of event A
!
Ac : non-occurrence of event A (also called as A-
complement)
!
A[B : either A or B has occurred
!
A\B : both A and B has occurred
!
⌦ : certain event, also called the universal set
!
⌦c = : impossible event, also referred as the null set

• If
A \ B = , then A and B are disjoint sets, i.e., A and B are mutually
exclusive events.
!
• A = B if A ⇢ B and B ⇢ A

10
Introduction to Probability

The probability of an event A is denoted by P (A) and is a number

nA
P (A) ⇡ N

where,
N is the number of experiment trials and
!
nA is the number of occurrences of event A
provided that N is sufficiently large.

The above definition is imprecise as it is not clear what is meant by

!
• approximately ⇡
!
• ``sufficiently large''
11
If we actually perform the experiment of tossing a fair coin N =10 times
and record the number of Heads and Tails, a sample realization could be
{ H T T T H T T T T T }.
!
Thus, n(H) = 2 and n(T ) = 8
! 2 8
P (H) = = 0.2, P (T ) = = 0.8.
10 10

For N = 100, n(H) = 57 & n(T ) = 43

!
P (H) = 0.57, P (T ) = 0.43.

When N = 1000, n(H) = 501 & n(T ) = 499 . This gives

!
P (H) = 0.501, P (T ) = 0.499.

12
The coin tossing experiment and estimating the probability of occurrences
of the outcomes reveals the following important points:

• The probability can be estimated only AFTER performing the

experiments. Hence, this definition of probability is referred as
the posteriori definition.
!
• Since the coin that is being tossed is fair, we expect
!
P (H) = P (T ) = 0.5
We see that for small values of N, the estimates are erratic and
not acceptable. However, as N is increased, the estimates
stabilize and for N=1000, the estimated probabilities are
approximately equal to the expected probability levels.
!
This highlights the importance of ``sufficiently large'' number of
samples.

13
Inductive Definition of Probability

The classical definition for probability is given by

NA
P (A) =
N

where
!
N are the possible outcomes of the experiment (total no. of sample points
in the sample space), and
!
NA no. of outcomes that are favorable for A (no. of sample points in ⌦A )

Note that the above definition is different from the previous one in that
here N no longer means trial runs of the experiment.

14
To explain the inductive definition more clearly, we consider the following
problems.

Example 1: Probability of getting an even number when rolling a fair dice.

Define the event A as getting even numbers.

!
! ⌦A = {2, 4, 6}, n(⌦A ) = 3
!
! ⌦ = {1, 2, 3, 4, 5, 6}, n(⌦) = 6
!
!
Therefore, P (A) = n(⌦ A)
n(⌦) = 3
6 = 0.5
!

15
Example 2: We roll two fair dice.
Find the probability of the sum of the numbers being 7.

In the next few slides, we apply the inductive definition to calculate the
probability.
!
In doing so, we present three possible solutions, of which two are incorrect.
!
We examine which is the correct solution and the corresponding reasons.

16
Solution 1:

We define the sample space of the sum of the numbers.

!
Sample space is thus given by

⌦ = {2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12}, n(⌦) = 11.

We define the event A as when the sum of the numbers on the faces of
the dice to be 7.
!
Thus, the sample space for A. is
! ⌦A = {7}, n(⌦A ) = 1.
!
Thus,
n(⌦A ) 1
P (A) = n(⌦) = 11 .

17
Solution 1 is WRONG !!
!
The solution is wrong because implicit in the solution is the assumption
that the probability of occurrence of each of the possible outcomes
listed in ⌦ have equal probability.
!
In reality, this is not true.
!
Thus, for example, the sum 2 can appear in only one way, viz, 1 on dice
1 and 1 on dice 2.
!
On the other hand, the sum say 7 can occur on any of the combinations
listed below:
!
! Dice 1 2 5 3 4 1 6
! Dice 2 5 2 4 3 6 1
!
! Sum 7 7 7 7 7 7
!
In the next solution, we take into account these distinct possibilities.
18
Solution 2:

We define the sample space for A in terms of (a, b) where a and b are
respectively the numbers that appear on the two dice.

Sample space is thus given by

!
⌦A = {(1, 6), (2, 5), (3, 4)}, n(A) = 3
!
while
!
⌦ = {(1, 1), (1, 2), (1, 3), (1, 4), (1, 5), (1, 6), (2, 2), (2, 3), (2, 4), (2, 5),
!
(2, 6), (3, 3), (3, 4), (3, 5), (3, 6), (4, 4), (4, 5), (4, 6), (5, 5), (5, 6), (6, 6)}
!
Thus,
! n(⌦) = 21.
!
The corresponding probability can thus be expressed as
!
n(⌦A ) 3
P (A) = n(⌦) = 21 .

19
Solution 2 is also WRONG !!
!
This is because here we do not differentiate between the outcomes of
the two dies. Thus, an outcome where we have 2 in the first dice and 5
in the second dice is considered to be identical when we have 5 in the
first dice and 2 in the second.
!
The outcome of not differentiating between the above two instances is
that once again we are making the assumption that the outcomes listed
in the sample space ⌦ have equal probability of occurrence when
actually this is not true.
!
Thus, for example, the combination (1,1) can happen only in one way
but the combination of 2 and 5 can happen in 2 ways and hence the
probability of occurrence of 2 and 5 is different than the probability of
occurrence of 1 and 1.

20
Solution 3:

Here, we differentiate between all the outcomes. Thus,

⌦ = {(1, 1), (1, 2), (1, 3), (1, 4), (1, 5), (1, 6),
(2, 1), (2, 2), (2, 3), (2, 4), (2, 5), (2, 6),
(3, 1), (3, 2), (3, 3), (3, 4), (3, 5), (3, 6),
(4, 1), (4, 2), (4, 3), (4, 4), (4, 5), (4, 6),
(5, 1), (5, 2), (5, 3), (5, 4), (5, 5), (5, 6),
(6, 1), (6, 2), (6, 3), (6, 4), (6, 5), (6, 6)}

Thus n(⌦) = 6 ⇥ 6 = 36
and the sample space for event A is given by
⌦A = {(1, 6), (2, 5), (3, 4), (4, 3), (5, 2), (6, 1)} n(⌦A ) = 6
n(⌦A ) 6
P (A) = n(⌦) = 36

This is the correct solution !!

Here, it is assumed that all the outcomes defined in the sample
space ⌦ are equally likely.
21