Module 11: The Uniform Distribution, the Normal Distribution, and the
Exponential Distribution
Much like with discrete distributions, there are certain classical continuous distributions well
worth understanding.
Module Objectives
Upon completion of this lesson, you should be able to:
1. Identify the uniform distribution
2. Work problems with uniform distributions
3. Identify the normal distribution
4. Work with the standard normal distribution
5. Identify the exponential distribution
6. Solve probability problems with the exponential distribution
Key Terms
uniform random variable standard normal histogram
De Moivre-Laplace distribution normal distribution
theorem interarrival times memoryless
uniformly distributed
Reading Assignment
Ghahramani, Chapter 07: Sections 7.1, 7.2 and 7.3
Module Introduction
The idea of drawing a point randomly from an interval provides the motivation for a simple, yet
useful random variable. Let X be the value of a random point selected from the interval (a, b).
Then X is called the uniform random variable over (a, b). The probability density function for
the uniformly distributed random variable X will be
We can simply integrate to get the density:
1
The graph of the density function is zero except for a step of height 𝑏−𝑎 on the interval (a, b). The
graph of the distribution function is zero to the left of a, a segment sloping from zero to height
one over the interval (a, b), and a horizontal line of height one thereafter. The textbook verifies
the uniform random variable has an expectation and variance given by
Example 1: The hardness of a sample of steel H (measured on the Rockwell scale) is a
continuous random variable uniformly distributed over the interval [50, 70]. What is the
probability that a sample is at least of hardness 63?
Definition A random variable X has the standard normal distribution if:
Unlike the uniform distribution, this integral is difficult; we’ll generally use the chart in the back
of the book for this. The density function of X will be:
The graph of f(x) is a bell shaped curve which is symmetric about the y axis. Bear in mind that
the integral associated with the standard normal distribution is non-elementary. The integrals
must be evaluated using tables or machines.
Let X have the standard normal distribution. The textbook verifies that E(X) = 0 and Var(X) = 1.
The student should note and memorize the continuity corrections given in the book for
approximating the binomial distribution by the normal distribution.
Example 2: Suppose the probability that patients are late for their appointments at a clinic is
historically 41 percent. If there are 100 patients scheduled, what is the probability that 40 of them
are late? Let X represent the number of late patients. The binomial distribution applies and we
can see that:
To use the approximation theorem, to approximate this calculation, we’ll need to find the mean
and standard deviation:
𝐸𝑋 = 𝑛𝑝 = 100 ∗ .41 = 41.
𝜎 = √𝑛𝑝(1 − 𝑝) = √100 ∗ .41 ∗ .59 ≈ 4.918
Using these values and continuity correction we get:
Definition A random variable X is called normal, with parameters μ and σ, if its density function
Is:
This random variable is denoted 𝑋~𝑁(𝜇, 𝜎 2 ); note carefully how the mean and variance
completely describe the normal random variable.
Our table in the back of the book assumes a mean of zero and a variance of 1. The following
lemma justifies the use of this table for all normal variable calculations:
Example 3: The gasoline sold per day at a local station is normally distributed with an expected
value of 1,500 gallons and a standard deviation of 132 gallons. If sales are less than 1,250
gallons, the station loses money for that day. What is the probability that the station loses money
no more than once a week? Let p be the probability that the station sells less than 1,250 gallons
of gas in a day. Let X be the normal random variable with mean 1,500 and standard deviation
132. Then:
The probability that no more than one day per week results in a loss is approximately:
The last of the special continuous distributions we will study is called the exponential
distribution. Let {N(t) : t ≥ 0} be a Poisson process. Recall that this represents the number of
observable events in the time interval [0, t]. Let Xk be the elapsed time between the arrival of the
(k - 1)st event and the kth event. X1, X2, X3, . . . is called the sequence of interarrival times of
the Poisson process. If 𝜆 = 𝐸(𝑁(1)), then
This produces the exponential distribution, defined as:
Differentiation leads to the density function:
The density function is said to be exponential with parameter λ. The textbook verifies the mean
and variance:
Example 4: The duration of phone calls is an exponential random variable with expectation of 5
1
minutes. Thus 𝜆 = 5, or λ=1/5. If someone gets to the phone just before you, then what is the
probability you wait more than 6 minutes to use the phone? Let X be the duration of the other
6
person’s phone call. Then we have P(X > 6) = 1 – F(6) = 𝑒 −5 .
An important property of the exponential distribution is the memoryless nature of the
distribution. That is, if X is an exponentially distributed random variable, then P(X > s + t | X > t)
= P(X > s), for all s, t >0. If you view X as the lifetime of an object, the preceding statement
asserts that if the object is functioning at time t, its remaining life has the same distribution as the
original. The object does not suffer wear and tear. It does not realize it has already been in
service for t amount of time. The memoryless property is equivalent to
P(X > s + t) = P(X > s)P(X > t).
Example 5: There are three tellers at the local bank. Each teller is servicing one customer, and
you are next in line. You know that your service will begin as soon as one of the customers
presently being serviced leaves the teller. If the amount of time a teller spends with a customer is
exponentially distributed with average time of 2 minutes, what is the probability that of the four
customers you will leave the bank last? When any one of the customers presently with a teller
leaves, you begin to be serviced. Since the exponential distribution is memoryless, the amount of
time the other two customers spend with their tellers will still be exponentially distributed with
the same parameter. For the other two customers, it will be as if their service had just begun.
Thus, by symmetry, the probability you leave last will be 1/3.
Module Assignment
Work the following problems from the textbook and submit them to the M11 Assignment in
Moodle.
Ghahramani, Section 7.1 p. 266–67, problems 1–4, 7, 10
Section 7.2 p. 281–83, problems 1–5, 9, 10, 12, 13, 16, 21, 27
Section 7.3 p. 290–91, problems 1, 3, 5, 7, 9, 10
Review p. 308–309, problems 1–5