Continuous RV Probability

Distributions
ENGDAT1 Lecture
Eric A. Siy
Dept of Industrial Engineering
De La Salle University

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Commonly Studied Patterns of Probability in
Continuous Random Variables
• Continuous Uniform distribution
• Normal Distribution
• Exponential Distribution
• Weibull Distribution
• Erlang Distribution

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Continuous Uniform Distribution
• Occurs if each outcome in an experiment (or observation) has equal
probability of occurrence over an interval (a<X<b)
Ex: arrival time of plane trips.
Ex: Exact age of children born in a given year. 1/365 probability

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Continuous Uniform Distribution
Parameters: a = Lower limit;
b = upper limit

Probability Distribution Function :

Cumulative Distribution Function:

Mean E(X) = Variance =

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Continuous Uniform Distribution
The heights of children in a certain section is uniformly distributed
between 53 to 62 inches. If a classmate is randomly chosen from
this class, what is the probability that his/her height is
a. at most 55 inches?
b. Between 56 to 60 inches?
c. At least 60 inches?

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Normal Distribution
• This is the most commonly observed
and assumed distribution in nature.
• Heights of adult men in a
population
• Anthropometric data (body
dimensions)
• The diameter of the hole made by
a drill press
• The score on a test
• The yield of grain for a plot of
ground
• The length of a newborn child

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Normal Distribution
• First described de Moivre in
x 

2 1733
1  12  
f ( x)  e    • Discovered by Carl Gauss in
2  1809.
• Elaborated in 1812 by Laplace
• Describes some natural
phenomena
• Describes sampling
characteristics of totals and
means

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Normal Distribution
Normal pdfs have two parameters
μ - expected value (mean “mu”)
σ - standard deviation (sigma)

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Two types of means and standard deviations

• The mean and standard deviation from

the pdf (denoted μ and σ) are
parameters
• The mean and standard deviation from
a sample (“xbar” and s) are statistics
• Statistics and parameters are related,
but are not the same thing!

Siy, Continuous Probability Distributions, DLSU Industrial

Engineering
Example: 68-95-99.7 Rule
• 68% of scores within
Wechsler adult intelligence scores: μ±σ
Normally distributed with μ = 100 = 100 ± 15
and σ = 15; X ~ N(100, 15) = 85 to 115
• 95% of scores within
μ ± 2σ
= 100 ± (2)(15)
= 70 to 130
• 99.7% of scores in
μ ± 3σ =
100 ± (3)(15)
= 55 to 145

Siy, Continuous Probability Distributions, DLSU Industrial

Engineering
Symmetry in the Tails
Because the Normal curve is symmetrical and the
total area under the curve is exactly 1…

… we can easily
determine the AUC in
95%
tails

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Determining Normal Probabilities
When value do not fall directly on σ landmarks:

1. State the problem

2. Standardize the value(s) (z score)
3. Sketch, label, and shade the curve
4. Use Normal Distribution Tables to find probabilities
from Z-scores.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Step 1: State the Problem
• What percentage of gestations (pregnancies) are less
than 40 weeks?
• Let X ≡ gestational length
We know from prior research:
X ~ N(39, 2) weeks

Find: Pr(X ≤ 40) = ?

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Step 2: Standardize
• Standard Normal
variable ≡ “Z” ≡ a
Normal random variable
with μ = 0 and σ = 1,
• Z ~ N(0,1)
• Use Normal Tables to
look up cumulative
probabilities for Z

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Example: A Z variable
of 1.96 has cumulative
probability 0.9750.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Step 2 (cont.)
Turn value into z score:
x
z

z-score = no. of σ-units above (positive z) or below
(negative z) distribution mean μ

For example, the value 40 from X ~ N (39,2) has

40  39
z  0 .5
2
Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Steps 3 & 4: Sketch & Lookup Normal Z-
Table
3. Sketch
4. Use Normal Tables to lookup Pr(Z ≤ 0.5) =
0.6915

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Probabilities Between Points
a represents a lower boundary
b represents an upper boundary
Pr(a ≤ Z ≤ b) = Pr(Z ≤ b) − Pr(Z ≤ a)

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Between Two Points
Pr(-2 ≤ Z ≤ 0.5) = Pr(Z ≤ 0.5) − Pr(Z ≤ -2)
.6687 = .6915 − .0228

.6687 .6915
.0228
-2 0.5 0.5 -2

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Values Corresponding to Normal Probabilities
1. State the problem
2. Find Z-score corresponding to percentile (Normal
Tables)
3. Sketch
4. Unstandardize:

x    z p
Siy, Continuous Probability Distributions, DLSU Industrial Engineering
z percentiles
 zp ≡ the Normal z variable with cumulative
probability p
 Use Normal Z-Tables to look up the value of zp
 Look inside the table for the closest cumulative
probability entry
 Trace the z score to row and column

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

e.g., What is the 97.5th
percentile on the Standard
Normal curve?
z.975 = 1.96

Notation: Let zp
represents the z score
with cumulative
probability p,
e.g., z.975 = 1.96
Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Step 1: State Problem
Question: What gestational length is smaller than 97.5% of gestations?
• Let X represent gestations length
• We know from prior research that
X ~ N(39, 2)
• A value that is smaller than .975 of gestations has a cumulative
probability of.025

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Step 2 (z percentile)
Less than 97.5%
(right tail) = greater
than 2.5% (left tail)

z lookup:
z.025 = −1.96
z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09

–1.9 .0287 .0281 .0274 .0268 .0262 .0256 .0250 .0244 .0239 .0233
Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Unstandardize and sketch
x    z p  39  (1.96)( 2)  35

The 2.5th percentile is 35 weeks

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
 Assessing Departures from Normality
Approximately Same distribution on Normal
“Probability Q-Q” Plot
Normal histogram

Normal distributions adhere to diagonal line on Q-Q plot

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Negative Skew

Negative skew shows upward curve on Q-Q plot

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Positive Skew

Positive skew shows downward curve on Q-Q plot

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Leptokurtotic

Leptokurtotic distribution show S-shape on Q-Q plot

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
• The time that a worker can finish a certain assembly task in a factory
follows a normal distribution with mean m=12.5 mins and standard
deviation s=1.8 mins. A new worker was asked to perform the
assembly task, and was timed at 14.2 minutes. What percentage of
worker population would be expected to finish the task at longer than
14.2 minutes?

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• The time that a worker can finish a certain assembly task in a factory
follows a normal distribution with mean m=12.5 mins and standard
deviation s=1.8 mins. What observed time would only the fastest 5%
of workers be able to achieve?

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Exponential Distribution

• If it can be shown that the number of arrivals during an interval is

Poisson distributed (i.e., the arrival times are Poisson distributed), then
the interarrival times are exponentially distributed
• The mean arrival rate is given by 1/b and the mean interarrival time is
given by b
• The Poisson distribution is a discrete distribution modelling
occurrences of arrivals.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Exponential Distribution

• It can be shown for the exponential distribution that the mean is equal
to the standard deviation; i.e., μ = b= σ
• The exponential distribution is the only continuous distribution that is
"memoryless", in the sense that P(X > a+b | X > a) = P(X > b)

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Exponential Distribution
• Parameter: b= expected value of X
• Probability Function:
• Cumulative Distribution Function: P(X≤x) = F(x) = 1-

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• The time a clerk spends to serve a customer is said to be exponentially distributed
with an expected time of 8 minutes per customer. Find the probability that a clerk
spends four to five minutes with a randomly selected customer.
The curve is:  f(x) = 1/8 e–(x/8)
Find P(4 < x < 5).
Solution:
The cumulative distribution function (CDF) gives the area to the
left. P(x < x) = 1 – e–x/8
P(x < 5) = F(5)= 1 – e(–5/8) = 0.7135 and 
P(x < 4) = F(4) = 1 – e(–4/8) = 0.6321

P(4<X<5) = F(5) – F(4) = 0.7135 – 0.6321 = 0.0814

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Weibull Distribution
• Weibull distribution can attain many shapes for various values of
shape parameter β.

• It can model a great variety of data and life characteristics, including

constant, increasing, and decreasing failure rates.

• Therefore, it is one of the most widely used lifetime distributions in

reliability engineering.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Weibull Distribution
• Parameters: β = shape parameter, or slope parameter
θ = scale parameter, or characteristic life parameter
 1
t   t  
• Probability distribution function: f (t )   
  
exp     
    

• Cumulative Distribution Function:   t  

F (t )  1  exp     
    
Note: if the shape parameter b=1, then the Weibull distribution is the Exponential Distribution.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

  1
t   t  
Shapes of the Weibull distribution f (t )   
  
exp     
    
β = shape parameter, or slope parameter
θ = scale parameter, or characteristic life parameter

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 • If q is increased while b is kept the same,
the distribution gets stretched out to the
right and its height decreases, while
maintaining its shape and location.
• If q is decreased while b is kept the
same, the distribution gets pushed in
towards the left (i.e., towards its
beginning or towards 0 ), and its height
increases.
• q has the same units as t, such as hours,
miles, cycles, actuations, etc.

http://reliawiki.org/index.php/The_Weibull_Distribution

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• An LED bulb of a certain brandname has a lifetime that follows a
Weibull Distribution with shape parameter b=2.5 and average
expected lifetime of q=18,000 hours. Determine the probability that
a randomly selected LED bulb of this brand would fail within the first
10,000 hours of operation.

 Pt   
F (t )  1  exp     
    

= 0.2055

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• An LED bulb of a certain brandname has a lifetime that follows a
Weibull Distribution with shape parameter b=2.5 and average
expected lifetime of q=18,000 hours. What is the probability that
this LED bulb will survive up to 25,000 hours?

  t  
F (t )  1  exp   PF(25,000)
 
    
=

= 0.1029653401

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Erlang Distribution
• A Danish mathematician named Agner Krarup Erlang set out to solve the key
problems in telephone network design in 1908 when he joined the
Copenhagen Telephone company. He wanted to determine how many
trunks are needed to carry a given amount of calling. His work would
collectively be a new branch of applied mathematics called Queuing Theory.

• His work was enormously influential worldwide such that in 1946, the
International Consultative Committee on Telephone and Telegraphs honored
him by adopting the name “Erlang”for the basic unit of telephone traffic.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Erlang Distribution

• An erlang is a dimensionless unit, referring to the continuous use of a

circuit expressed as the number of calls-minutes received by a telephone
hub in an hour (60 minutes). One erlang is 60 minutes’ worth of calls in
60 minutes. 10 erlangs means 600 minutes worth of calls was received
in 60 minutes, like , say, receiving 200 calls of 3 minutes’ duration per
call (200x3 call-minutes/60 minutes).

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Erlang Distribution

The Erlang distribution is used to describe the total time that a series
of exponentially distributed task times would take. Say, if two tasks 1 and
2 are the series of tasks that an operator needs to do to service a call, and
task 1 can be done in t1 seconds, with an exponentially distributed
expected time, and task 2 in t2 seconds, then the total time that these
two tasks would take together (t= t1+t2) is not exponentially distributed
but rather follows an Erlang distribution.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Erlang Distribution
• Parameters: b=positive real number
(expected time for all tasks or calls)
k =positive integer (number of tasks or trunks)
k
 1

k  
• Probability distribution function: f (t )  
k
 t k 1e 
t

k  1 !
When k=1, the Erlang distribution becomes the exponential distribution.
 
•Expected value: E(x) = b

•Variance: Var(x) = b2/k

Siy, Continuous Probability Distributions, DLSU Industrial Engineering
Some notes about the Erlang Distribution
• Since the Erlang is a continuous variable distribution, then integration
of the function must be performed to find the probabilities that a
variable t would be within certain intervals of values. As such, our
discussion will not contain any examples for this. What we will use
instead is a particular formula called Erlang B that has been found
useful in industry.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

 Some notes about the Erlang Distribution

• An important application of
Erlang Distributed variables is in
finding the probability that a
given caller at a telephone
network would find that all the ….gives the probability that the network is busy,
that is, all n (parallel) trunk server lines are busy
available lines are already full. servicing some call signal.
We call this applied formula the Given that the network has
Erlang B formula: A = erlangs of call traffic.
N= number of trunk lines

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• A village has a telephone exchange center where all calls going out of the
village and into the village must pass through. The village has N=8 lines,
while there is an expected A=5 erlangs call traffic (ex. 150 calls per hour
with each call taking 2 minutes each= 150x2/60= 5 erlangs.) What is the
probability that a randomly selected caller would find the telephone
exchange network with no lines available for a call (i.e. network is busy)?
Given A=5 erlangs, N=8 lines.

There is a 7% chance that a caller would not be able to call out or into the
village.

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• According to www.tarrani.net/linda/ErlangBandC.pdf
Telemanagement # 187, a circular distributed to telecommunication
professionals, the following evaluations on the values of P is generally
accepted:

Target p=0.05 as acceptable. P=0.10 is terrible.

P=0.001 is so good that most callers will never get a busy signal.

Further Erlang tutorials can be found in:

http://plus.maths.org/issue2/dar/index.html
http://www.erlang.com

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

Practice Exercises (Continuous Probability distributions)

1. A research scientist reports that mice will live an average of 40 months when their diets are sharply
restricted and then enriched with vitamins and proteins. Assuming that the lifetimes of such mice are
normally distributed with a standard deviation of 6.3 months, find the probability that a given mouse
will live:
a. More than 32 months
b. Less than 28 months
c. Between 37 and 49 months

2. If the length of a Filipino’s forefinger averages 4.5 inches following a normal distribution with a standard
deviation of 0.75 inches, what is the length of the longest 1% (or top 1%) of the population?

3. The life of a small motor is normally distributed with mean of 10 years with standard deviation of 2 years.
a. What is the probability that the motor will function after 11.5 years?
b. What is the probability that the motor will function between 9.2 and 9.6 years?
c. The manufacturer replaces free all motors that fail while under guarantee. If he is willing to replace
only 3% of the motors that fail, how long (years) a guarantee should he offer?

Siy, Continuous Probability Distributions, DLSU Industrial Engineering

• Practice Exercises (Continuous Probability distributions)

4. It is known that a certain measurement is normally distributed with mean 1.50 cms and standard
deviation 0.2 cm.
a. Find the measurement above which lies the biggest 4.01% of all components.
b. Gauges are used to reject all components where a certain dimension is not within the specification 1.5
+ k. Determine the value of k such that the specifications cover 95% of the measurements.
c. Using the normal distribution, what is the approximate probability that at least 25 of 1000 of these
components have dimensions of more than 1.856?
5. The mean lifetime of a calculator battery can be modeled with an exponential distribution with a mean
lifetime of 3 years. What is the probability that a newly installed battery would last at least 4 years?
Between 2 to 3.5 years?
6. The life of a Nokia cellphone battery for the 9980 model has an exponentially distributed lifetime with
mean lifetime of 1000 hours. What is the probability that a certain battery of this type would last more
than 1200 hours?
7. A call center receives calls at a rate of 300 call-minutes per hour. It currently has 12 trunklines that can
accommodate these calls. What is the expected probability that a caller would get a busy signal when
he/she calls the center?
8. Calculate the number of trunk lines that must be installed to accommodate a traffic rate of 15 erlangs
with a probability of 0.02 of being blocked (or busy signal).

Siy, Continuous Probability Distributions, DLSU Industrial Engineering