Probabilistic Methods in Engineering

Exercise Set 4
Date Due: 3:50 PM, Thursday, the 19th of March 2009

Office hours: Thursdays, 12:30-1:30 PM and on the SAKAI system

You are required to compose your solutions in neat and legible handwriting. Up to 10% of the total score may
be deducted solely due to the apearance and legibility of your writing and your use of the English language.
In order to obtain the highest possible score, make sure that you explain your reasoning. Often, simple formulae
are not enough to answer a question. Explain what you are doing! This will also ensure that you get a large
fraction of the total points even if you make a mistake in your calculations. In short write simple, whole
grammatical sentences that include a subject, verb and object.

Exercise 1. Liquid products were first obtained from coal in England during the 1700s. Lamp oil was produced
from coal in the United States as early as 1850, but the US coal chemicals industry did not develop until World
War I. A modern coal-for-recovery system uses a battery of coke ovens to produce liquid products from the
coal feed. These observations are obtained on the random variable X, the number of gallons of liquid product
obtained per ton of coal feed.
7.6 6.1 7.4 8.8 8.7 9.0 7.6 7.7 8.2 6.2 9.5 9.6 7.8 8.8 6.7 9.1
7.1 7.6 9.2 9.7 8.7 7.3 8.1 7.9 10.0 6.2 8.0 6.8 8.2 7.9 6.2 8.7
6.5 9.5 8.5 7.1 8.2 7.1 5.3 8.4 9.6 6.7 9.3 7.7 7.4 7.9 7.4 8.1

i) Construct a stem-and-leaf diagram for these data.

ii) Is the assumption that X is normally distributed justifiable? Explain.
iii) Divide the data into six categories using the method described for histograms in the textbook. Construct
a relative frequency histogram of the data. Does the histogram exhibit the bell-shape characteristic of a
normal density?
iv) Construct a relative cumulative frequency ogive for these data. Use the ogive to approximate the proba-
bility that a randomly selected ton of coal will yield less than 7 gallons of liquid product.

(2 + 1 + 3 + 3 Marks)
Exercise 2. Temperature differences between the warm upper surface of the ocean and the colder deeper levels
can be utilized to convert thermal energy to mechanical energy. This mechanicl energy can in turn be used
to to poduce electrical power using a vapor turbine. Let X denote the difference in temperature between the
surface of the water and the water at a depth of 1 kilometer. Measurements are taken at 15 randomly selected
sites in the Gulf of Mexico. These data result in the following temperatures:
22.5 23.8 23.2 22.8 10.1∗ 23.5 24.0 23.2 24.2 24.3 23.3 23.4 23.0 23.5 22.8

i) Construct a double stem-and-leaf diagram for these data.

ii) Find the sample mean, sample median and sample standard deviation for these data.
 iii) Note that the starred observation in the data set is very different from the others. It is a potential outlier.
Construct a boxplot for these data to verify that the value 10.1 does, in fact, qualify as an outlier.
iv) To see the effect of this outlier, drop it from the data set and calculate the sample mean, median and
standard deviation for the remaining 14 observations. What measure is least affected by the presence of
the outlier? Do you see why it is desirable to report both the mean and the median of a data set?

(1 + 1 12 + 2 + 2 12 Marks)
Exercise 3. An interactive computer system is available at a large installation. Let X denote the number of
requests for this system received per hour. Assume that X has a Poisson distribution with parameter λs. These
data are obtained:
25 20 20 30 24 15 10 23 4

i) Find an unbiased estimate for λs.

ii) Find an unbiased estimate for the average number of requests received per hour.
iii) Find an unbiased estimate for the average number of requests received per quarter hour.

(2 + 2 + 2 Marks)
Exercise 4. Consider the experiment of rolling a pair of dice until a sum of 7 is obtained. let X denote the
number of trials needed to obtain a sum of 7.
i) What is the distribution of X?
ii) What are the theoretical average µ and variance σ 2 of X?
iii) Perform the expriment described 25 times, and thus obtain a sample of sixe n = 25 observations on X.
Plot a stem-and-leaf diagram for your data. Does the distribution appear to be symmetric? Use your data
to obtain unbiased extimates for µ and σ 2 . Compare your answers to the true values of these parameters.
iv) Consider the random variable X, the average number of trials needed to roll a sum of 7 based on 25 trials.
What is E[X]? What is Var X?
v) Obtain the observations of X from 20 of your classmates. Plot these values on a number line. Do they
fluctuate about µ as epected? Find the average value of the observed X values. Is it close to µ as expected?
Find the sample variance of the X values. Is this sample variance close in value to σ 2 /25 as expected?
(1 + 1 + 3 + 1 + 4 Marks)
∑n
Exercise 5. In this exercise you will show that the naively exected estimator for σ 2 , namely, i=1 (Xi − X)2 /n
is a biased extimator for σ 2 and tends to underestimate the true variance. Let X1 , X2 , . . . , Xn be a random
sample of size n from a distribution with mean µ and variance σ 2 .
∑n
i) Show that i=1 (Xi − X)2 /n = (n − 1)S 2 /n.
∑n
ii) Verify that E[ i=1 (Xi − X)2 /n] < σ 2 .
∑n
iii) Compare the value for i=1 (Xi − X)2 /n obtained from your data in Exercise 4 iii) with the true variance
σ 2 of Exercise 4 ii).
(1 + 1 + 2 Marks)
Exercise 6. Let W be an exponential random variable with parameter β unknown.
i) Find the method-of-moments estimator for β based on a sample size n.
ii) Find the maximum-likelihood estimator for β based on a sample size n.
(1 + 1 Marks)
Exercise 7. Let X be an binomial random variable with known parameter n and unknown parameter p.

i) Find the method-of-moments estimator for p based on a sample size m.

ii) Find the maximum-likelihood estimator for p based on a sample size m.

(1 + 1 Marks)
Exercise 8. A new material is being tested for possible use in the brake shoes of automobiles. These shoes
are expected to last for at least 75,000 miles. Fifteen sets of four of these experimental shoes are subjected
to accelerated life testing. The random variable X, the number of shoes in each group of 4 that fail early, is
assumed to be binomially distributed with n = 4 and p unknown.
i) Find the maximum-likelihood estimate for p based on these data:

1 0 1 0 2 1 0 1
0 1 0 0 0 1 0

ii) If an early failure rate in excess of 10% is unacceptable from a business point of view, would you have
some doubts concerning the use of this new material? Explain.
(2 + 1 Marks)

Exercise 9. [This exercise is taken from the second midterm exam of Fall 2008.] Suppose that the random
variable X has the probability distribution
{
(γ + 1)xγ , 0 < x < 1,
f (x) =
0, otherwise.

Let X1 , X2 , . . . , Xn be a random sample of size n. Find the maximum likelihood estimator of γ.

(2 Marks)

Exercise 10. Prove the following theorem:

Let X1 , . . . , Xn be a random sample of size n from a normal distribution with mean µ and variance
σ 2 . Then X is normally distributed with mean µ and variance σ 2 /n.
The steps of the proof are outlined in Exercises 39-42 of Section 7.3 of the textbook (pages 251-252). Instead
of just giving the answers to these indvidual exercises, formulate a stand-alone, coherent proof of the theorem.
(If you wish, you may formulate and prove some preliminary lemmas on the moment-generating functions of
linear combinations of normal random variables.)
(5 Marks)