You are on page 1of 6

CE 307, Fall 2019 Name: ___________________________________

Exam II

Greetings! This here is the instructions. Briefly:


This test is open book / open note. Additionally, you are allowed to use a spreadsheet (Excel) and
access any course materials on Beachboard. You are not allowed to visit the internet at large, including
your email, any chat programs, search engines, websites, anything that isn’t our course material on our
site.
As I am allowing the use of a spreadsheet, you still need to show your work on this exam to get credit.
Showing your work can include the mathematical expression you solve or the spreadsheet function you
used. As long as I can figure out what you are trying to do, I can give you credit for it.
Don’t cheat. Cheating sucks. Cheaters never prosper. If you are looking to cheat I can tell you I might
not stop you, but what I do have are a very particular set of skills. Skills I have acquired over a very
long career. Skills that make me a nightmare for people like you. If you cheat, I will find you and I will
fail you. (Also as per the syllabus, your final grade is reduced by one whole letter.)

Page 1 of 6
CE 307, Fall 2019 Name: ___________________________________

Problem 1. Traffic sucks1 (20 points)


Traffic engineers have noted that, on average 1 out of every 8 cars (12.5%) turns left at a given
intersection. They are considering a redesign of the intersection to include a protected left turn lane
and want to consider some simple statistics. The flow of traffic toward the intersection can be modeled
as a normally distributed variable with a mean of 30 cars/min and a variance of 25 cars2/min2.
A) What is the 90th percentile of traffic flow approaching the intersection?
B) What is the probability that in any given minute 40 or more vehicles will approach the
intersection?
C) If 32 cars approach the intersection during a red light (which conveniently lasts exactly 1
minute), what is the probability that more than 6 will be turning left?
D) Assuming 35 cars/min approaching the intersection is the design traffic flow, how many
cars/min must the turn lane be designed to accommodate for, if the engineer designs for the 75th
percentile of left turners.

We have 2 different sets of statistics to consider here – traffic flow and left turners. The traffic
flow is a continuous normally distributed variable. Turning left is a binomial distribution (or if
you want to be fancy, you could use a poisson distribution).
a) P = 0.9 =NORM.INV(0.9,30,SQRT(25)) = 36.4 cars
b) P(X> 40) = 1 - P(X<40) =1-NORM.DIST(40,30,5,1) = 0.02275
c) P(X>6) = 1 – P(X<=6) =1-BINOM.DIST(6,32,0.125,1) = 0.096478
d) find value of x such that P = 75%. =BINOM.INV(35,0.125,0.75) = 6
(If you were fancy and tried poisson distribution, you fiddle with x until you get a value > 75%
=POISSON.DIST(6,35*1/8,1) . In both cases, 5 cars turn with about 70% frequency while 6 is
in the 80% range. We pick the value that’s greater than 75%, which is 6)

1 The values in this problem are completely made up. I don’t think I’ve ever actually taken a traffic class. I remember
designing a banked curve once for an assignment, but that’s about all I remember from the one transport class I took.

Page 2 of 6
CE 307, Fall 2019 Name: ___________________________________

Problem 2. Redman’s Red Sauce2 (25 points)


I’m tired of being a teacher and instead want to open a pizza joint. But being a nerd, I have all sorts of
fancy gadgets. One of my time saving tools in an automatic pizza sauce dispenser. A 12” pizza (the
only size I sell) requires 4 ounces of sauce3. The dispenser is set to dispense an expected volume of 4
ounces. The manufacturer indicates that the standard deviation is 10% of the set volume and that the
dispensed volume is normally distributed.
A) What is the probability that a given pizza has more than 4.5 ounces or less than 3.5 ounces of
sauce on it? (Tony D doesn’t like overly wet pizzas nor does he like his pizza dry.)
B) Suspecting something is wrong with my dispenser, I carefully measure the volume dispensed on
one pizza. It was 4.3 ounces. What is the probability of getting a dispensed volume 4.3 ounces
or greater?
C) Repeating the experiment in part B, I measure the dispensed volume on 10 pizzas. It averaged
4.35 ounces. What is the probability of getting an average dispensed volume of 4.35 ounces or
greater?
D) Each day, my pizza joint sells exactly 100 pizzas. How much sauce will be used per day?4
Express the answer as a mean and standard deviation.
E) Given the answer to part D, if I want to make enough sauce to cover the 95th percentile of sauce
usage, how much should I make?

a) P(X<3.5 or X>4.5) =NORM.DIST(3.5,4,0.4,1) + (1-NORM.DIST(4.5,4,0.4,1)) = 0.2113


b) P(X>4.3) = 1- P(X<4.3) =1-NORM.DIST(4.3,4,0.4,1) = 0.226627
c) now it’s an average of 10, so we use standard error
P(Xbar > 4.35) = 1- P(Xbar <4.35) =1-NORM.DIST(4.35,4,0.4/SQRT(10),1) = 0.002829
d) mu = 100*4 = 400
sd = =SQRT(100*0.4^2) = 4
e) =NORM.INV(0.95,400,4) = 406.5794 oz

2 That’s the name of the pizza joint I’m opening in this problem.
3 At least according to PMQ.com, which I assume stands for Pizza Magazine Quarterly
4 Do NOT convert into pounds or gallons or whatever. Leave it in ounces.

Page 3 of 6
CE 307, Fall 2019 Name: ___________________________________

Problem 3. Grading frustration. (20 points)


It’s practically a given that one third of the students taking statics will either drop or fail the course5.
When I grade midterms I get frustrated very easily. Let’s assume that there are 30 students in my class,
the midterm has 5 problems and that each student has a 20% chance of doing something stupid on
every problem. In this binary world, doing something stupid means they get the problem wrong; not
doing something stupid means they get the problem correct.
A) What is the probability that a student gets more than 3 problems correct on their test?
B) When grading tests, I grade by problem – grade all of the problem 1s, then all of the problem
2s, etc. What is the probability that 15 or fewer people get the first problem correct?
C) Every 10 problems I grade, I take a break. However, if I encounter more than 3 stupid
mistakes, I will instantly stop and take a break. What is the probability that I stop grading the
current group of 10 exams early.
D) Tricky part: I said that approximately one third of the class does not pass; this may not be
accurate given the above statistics. Let’s assume not passing means getting 3 or more wrong on
the test. A 20% chance of making a mistake does not equate to 1/3 of the class not passing.
a) What fraction of the class does not pass if they each have a 20% chance per problem of
making a mistake?
b) What chance of making a mistake would give an overall not pass rate of 1/3 of the class?

Right or wrong, constant probability = bernouli trial and all the associated distributions. Let’s
write the probability and it’s complement: p_w = 0.2; p_r =0.8. pick appropriate for each
problem
a) n = 5 problems, p = 0.8, P(X>3) = P(4 or 5) =BINOM.DIST.RANGE(5,0.8,4,5) = 0.73728
b) n = 30 tests, p = 0.8, P(X<=15) =BINOM.DIST(15,30,0.8,1) = 0.000231
c) this is tricky. IF we get 4 wrong answers, we stop, so that’s negative binomial. And we
want to the 4 wrong answers to occur before test 10 (because that’s a normal stop).
Switching to p = 0.2 (we call a wrong answer a success), we can have at most 5 failures with
4 successes. =NEGBINOM.DIST(5,4,0.2,1) = 0.085642
d) If p = 0.2, then you can make a simple table of # wrong or # right (depending on how you
calculated it) with a binomial distribution, n = 5 trials, p = 0.2 (or 0.8) and calculate the
probability distribution for all of the outcomes, x = 0 through 5. Then simply add up all of the

5 Probably because they don’t know the simplest of things – like the fact that a fixed connection resists a moment.

Page 4 of 6
CE 307, Fall 2019 Name: ___________________________________

probabilities for 3 or more wrong (which equates to failing). If your spreadsheet is set up
properly, then part b, then simply adjust the p value until you get a sum around 33%.
wrong right p wrong right p
5 0 0.00032 5 0 0.011586
4 1 0.0064 4 1 0.08336
3 2 0.0512 3 2 0.239914
2 3 0.2048 2 3 0.345242
1 4 0.4096 1 4 0.248406
0 5 0.32768 0 5 0.071492
p 0.41
fail 0.05792 fail 0.33486

Page 5 of 6
CE 307, Fall 2019 Name: ___________________________________

Problem 4. Rocket man! (15 points)


After scrapping the pizza joint, I decide to make rockets and see if I can do a better job than SpaceX.
I’ve developed prototype rockets that launch and then usually land back where they started from.
We’ve done enough testing that we have developed a distribution of the Δx – the distance from the
launch point that the rocket lands. This distribution is shown below:

Distribution of landing distances


f(Δx)

0 1 2 3 4 5 6
Δx, meters

Based on this distribution:


A) What is the mean distance from the launch point that RedmanY rockets land?
B) What is the variance of Δx?
C) What is the probability that the rocket lands within 1 meter of the launch point?

First, put a scale on the vertical axis. Recall that the continuous distribution has an area of 1. So find
height of triangle with area 1. 1 = 1/2*5*f => f = 0.4 @ x = 0
now write as an equation: f(x) = -0.4/5x + 0.4 = -0.08x + 0.4
5

a) mu = ∫ x f ( x dx)=∫ −0.08 x 2 +0.4 x dx=−0.08 /3 x3 +0.4 /2 x 2=1.6667


0

b) sigma^2 = ∫ x 2 f ( x) dx− μ 2=∫ −0.08 x 3 +0.4 x 2 dx− μ 2 =−0.08 / 4 x 4 +0.4 /3 x 3−1.6667 2=1.3889
0

c) f(x = 1) = 0.32. Area (1-5) = 0.5*0.32*4 = 0.64. within 1 m then, is 1 – 0.64 = 0.36
(the lazy among you will realize that A and C can be calculated somewhat easily. A is simply the
centroid, 5/3, and C is just area math, as shown. B, unfortunately, involves integration)

Page 6 of 6

You might also like