ENGI 9411

MODELING UNCERTAINTY
PART II
Normal or Gaussian Distribution
Normal distribution is denoted as, (, )
PDF
2
1 1
() = , < < + 1
2 2

CDF 2
1 1
() = 2
2 2

Simplifying by transforming into a standard normal variable with zero mean and unit standard deviation:

= 3

Equations(1) and (2) become

1 1 2
() = , < < + 4
2 2 2
Normal or Gaussian Distribution

3
Standard Normal or Gaussian Distribution

1 1
() = 2 5
2 2

The standard normal distribution is denoted as 0,1 and its CDF as that is,
= ()

= 1 = 6
Also
= 1 = 1 1 (7)

For example,
0.1 = 1 1 = 1 0.53983 = 0.46017 8

1 0.10 = 1 0.90 = 1.28 (9)

Furthermore,

< = (10)
4
Example 1
Suppose a steel cable has to carry a weight of 10 kips. Information on the strength of similar cables
indicates that the strength of the cable, R, can be modeled by a normal random variable with a mean
of 25 kips and a standard deviation of 5 kips. Calculate the probability that the cable will be unable to
carry the weight, or the probability that the cable will break.

Solution
= = 10

10 = ( < 10)
From equation (10),

10 25 25
= = 3
5 5

= 1 3 = 1 0.99865

= 0.00135
5
 Lognormal Distribution
Lognormal distribution considers the natural logarithm of a variable, X, which eliminates the possibility of
negative values. If a random variable has a lognormal distribution, then its natural logarithm has a normal
distribution. This is the meaning of the term lognormal. The PDF of a lognormal variable is unsymmetrical
and is given by:

2
1 1
() = , 0 < < + 11
2 2
The two parameters, , , calculated from the parameters of the normal distribution as:

1
= = 2 12
2
2
2

= Var lnX = ln 1 + = ln 1 + 2 (13)


When = = < 0.3,

6
Lognormal Distribution

7
Lognormal Distribution
Following the procedure for standard normal distribution for a lognormal variable, the standard variable, S,
is given by:

= 14

Such that

< = 15

If is the median of a lognormal variable, then

= ln (16)

= (17)
1+ 2

8
Beta Distribution
The beta distribution is a very flexible distribution used when a random variable is known to be bounded by
two limits, a and b.
PDF
1 1 1
= , 17
, +1
= 0, elsewhere
Where q and r are the parameters of the distribution and (, ) is the beta function.

=+ (18)
+
And
2
= 19
+ 2 ++1

, = 1 1 1 (20)
0 9
Beta Distribution
()()
, = 21
( + )

If a = 0 and b = 1, equation (17) simplifies to the Standard beta distribution ,Equation (22):

1
= 1 1 1
,0 1 22
,
= 0, elsewhere

When = = 1, the beta distribution becomes a uniform distribution.

10
Beta Distribution

11