You are on page 1of 5

Mathematics 370 Probability and Statistics I

November 9, 2014
Homework #10
ANSWER KEY

Online homework sample answers.


1. If a is uniformly distributed over [5, 13], what is the probability that the roots of the equation
x2 + ax + a + 3 = 0
are both real?
We require a non-negative discriminant: a2 4(a + 3) 0. The discriminant factors as (a 6)(a +
2) 0. Both factors are non-positive if a 2, and both are non-negative if a 6. Thus we want the
probability of the intervals [5, 2] and [6, 13]. The total probability is 3/18 + 7/18 = 0.5556.
2. A manager of a department store reports that the time of a customer on the second floor must
wait for the elevator has a uniform distribution ranging from 1 to 6 minutes. If it takes the elevator
30 seconds to go from floor to floor, find the probability that a hurried customer can reach the first
floor in less than 4.5 minutes after pushing the elevator button on the second floor.
The customer can only wait 4 minutes, since the additional 30 seconds is needed for travel time.
So P (X 4) = (4 1)/(6 1) = 0.6.
3. Suppose that the time (in hours) required to repair a machine is an exponentially distributed
random variable with parameter = 2.5. What is
(a) the probability that a repair time exceeds 3 hours?
(b) the conditional probability that a repair takes at least 13 hours, given that it takes more than
9 hours?

R

For (a), we simply compute 3 ey/2.5 /2.5 dy = ey/2.5 = e3/2.5 = 0.3012.
3

Pr(Y > 13)


(b) P r(Y > 13|Y > 9) =
, since {Y > 13} is contained in {Y > 9}.
P r(Y > 9)
R y/2.5
e
/2.5 dy
e13/2.5
So R13
= 9/2.5 = e4/2.5 = 0.2019.
y/2.5
/2.5 dy
e
9 e
4. Because of the relatively high interest rates, most consumers attempt to pay off their credit card
bills promptly. However, this is not always possible. An analysis of the amount of interest paid
monthly by a banks Visa cardholders reveals that the amount is normally distributed with a mean of
30 dollars and a standard deviation of 8 dollars.
A. What proportion of the banks Visa cardholders pay more than 33 dollars in interest?

B. What proportion of the banks Visa cardholders pay more than 40 dollars in interest?
C. What proportion of the banks Visa cardholders pay less than 18 dollars in interest?
D. What interest payment is exceeded by only 18% of the banks Visa cardholders?
Here the interest amount is X N (30, 82 ).
A. P r(X > 33) = P r(Z > (33 30)/8) = P r(Z > .38) = .3520.
B. P r(X > 40) = P r(Z > 10/8) = P r(Z > 1.25) = .1056.
C. P r(X < 18) = P r(Z < 12/8) = P r(Z < 1.5) = P r(Z > 1.5) = .0668.
D. We need to find x such that P r(X > x) = .18. For Z, the corresponding value is about .92. So
x = 30 + .92(8) = 37.36.
5. Womens weights are normally distributed with a mean given by = 143 lb and a standard deviation
given by = 29 lb. Find the seventh decile, D7 , which separates the bottom 70% from the top 30%.
For a standard normal, the seventh decile is about .52, as you find by looking for .30 in the normal
table. So the corresponding value for the weights is 143 + .52(29) = 158.08.
6. The probability density function of X, the lifetime of a certain type of device (measured in months),
is given by

0 if x 6
f (x) =
6
if x > 6
x2
Find the following: (i) P (X > 28) =
(ii) Thecumulative distribution function of X:
if x 6
F (x) =
if x > 6
(iii) The probability that at least one out of 4 devices of this type will function for at least 31
months.

R

(i)P r(X > 28) = 28 6/x2 dx = 6/x = 0.2143.
28

0
if x 6
(ii) F (x) =
1 6/x
if x > 6
The second formula is found by integrating from 0 to x.

R

(iii) First, P r(X > 31) = 31 = 6/x 0.1935. We need to assume independence of the devices
31
to proceed, but this is reasonable. Then the problem becomes a binomial distribution. We consider the
complementary probability, that none of them last that long. It is (1 .1935)4 = 0.4230. Consequently
the answer is 0.5770.
7. Your company is making a sealed bid for a construction project. If your firm has the lowest bid, you
will pay another firm (a subcontractor) 160 thousand dollars to do the work. If you believe that the

minimum bid (in thousands of dollars) of the other firms bidding on the project is a random variable
that is uniformly distributed on [130, 220], how much should you bid to maximize your expected profit?
Let B be the size of your bid. There is no purpose to bidding less than 160, so 160 B 220. If
the others minimum bid, call it X, is less than B, your profit is 0. If X > B, your profit is B 160.
Thus, the expected profit is 0 P (X < B) + (B 160)P (X > B) = (B 160) 220B
90 . So we want
to maximize the function B 2 + 380B 35200, where I have thrown the denominator away. Set the
derivative equal to zero: 0 = 2B +380. This gives B = 190. Clearly there is no profit in the endpoints
with B = 160 and B = 220, and 190 gives a positive profit, so this is the best.
8. Suppose Y Exp(7). Find the mean, median and modeZ of this random variable.

y y/7
The mean can by computed by integrating by parts on
e
dy, or you can just use the fact
7
0
that the parameter is the mean, namely 7.
The median is the point that is greater than (less than) 50% of the possible
R values, i.e. if ym is
the median, then P r(Y ym ) = .5. Then we must solve the equation .5 = ym 71 ey/7 dy = eym /7 .
Then ym = 7 ln .5 4.8520.
Finally, the mode is the high point of the density, but an exponential density is strictly decreasing.
Hence the left endpoint, y = 0 gives the mode.
9. For certain ore samples the proportion Y of impurities per sample is a random variable with density
function given by
 8 3
2 if 0 y 1
3y + y
f (y) =
0
otherwise
The dollar value of each sample is W = 3 1.75Y . Find the mean and variance of W .
I will find
1
Z 1 EY and V Y first. 5
8 3
8y
y 4
47
2
EY =
y( y + y ) dy =
+ = .
3
3 5
4 0 60
0
1
 6
Z 1
8y
y 5
29
2
2 8 3
2
EY =
y ( y + y ) dy =
+ = .
3
3
6
5
45
0
0
47 2

(
)
=
0.03083
So V Y = 29
45
60
Now EW = 3 1.75EY = 1.6292, and V W = (1.75)2 V Y = 0.09442.
10. A candy store owner has daily demand Y for a certain brand of candy sold from the bulk bins.
Suppose the owner will never stock more than one crateful of this candy (so that it will not go stale),
and that a crate can fill 50 bins. We measure Y in terms of fractions of a crate. With these units, Y
has density function


f (y) =

3y 2 if 0 y 1
0
otherwise

The grocer can buy a bins worth of candy for 0.40 dollars and sell a bins worth for 1.00 dollars.
What amount of candy, C, in bins, should the store owner purchase to maximize expected daily profit?
I will first work in fractions of a crate, and solve for the daily purchase amount c. My wholesale
price is $0.40, and my retail price is $1.00. It is OK to leave these as is in this example, without
converting to the cost per crate, because the conversion factors drop out during the computation
(everything is strictly linear with respect to these prices.) With these assumptions, profit will be the
difference between the revenue brought in (price times demand or available supply c, whichever is less)
and total cost (wholesale price times supply c.)
So if g is the profit function,
Z c
Z 1
2
g(c) =
(y .4c)3y dy +
(1 .4)c3y 2 dy
0
c
c
1
= .75y 4 .4cy 3 0 + .6cy 3 c
= .25c4 + .6c
Differentiate and set equal to zero: 0 = g 0 (c) = c3 + .6, so c = .8437. The number of bins is
50(.8437) = 42.1716.
11. Let X be a random variable with probability density function

c(8x x2 ) if 0 < x < 8
f (x) =
0
otherwise
Find the value of c:
c=
Find the cumulative distribution function of X:

if x 0

if 0 < x < 8
F (x) =

if x 8
R8
To find c, solve 1c = 0 8x x2 dx, which gives c = 3/256 = 0.01171875.
We know the formulas for the c.d.f. are trivial except on the support of X. We need the indefinite
integral on the support.

if x 0
0
2
3
3/256(8x /2 x /3) if 0 < x < 8
F (x) =

1
if x 8

12. The density function for a random variable X is given by



a + bx2 if 0 x 1
f (y) =
0
otherwise
If EX = 0.7, find a and b.
R1
Since we have a probability distribution, 0 a + bx2 dx = 1, which implies a + b/3 = 1.
R1
From the expectation, 0.7 = 0 x(a + bx2 ) dx = a/2 + b/4.
Solving the linear system gives a = 0.2, b = 2.4.
13. You are shopping for a condominium. You see an advertisement for a condo in Protractor Bend,
a subdivision with a very large number of condominiums,all spaced uniformly along a perfectly semicircular street that is attached to a perfectly straight, noisy freeway. The semicircle has radius 3
miles.
As a mathematics student, this intrigues you. (As an architect, it puzzles you.) You realize that
units that are farther away from the freeway will be more expensive, so the distance from the freeway
is an important consideration.
Let X measure the shortest distance (as the crow flies) from a randomly chosen condo in Protractor
Bend to the freeway.
Find the following: P (1.5 < X < 3) =
In my case, we measure from 1.5 miles back all the way to the apex of the semicircle, and have
a single arc. Since there is a uniform distribution on the semicircle, I just need to figure out what
fraction of the total length this is. If you center the coordinate system in the middle of the freeway
beneath the apex of the semicircle, then 1.5/3 = .5 is the sine of the angle that intersects the right
edge of my arc, and the arcsine is /6. The angle for the whole semicircle is , so my arc corresponds
to 2(/6) = 2/3. This is 2/3 of the arc, so that is also my probability.
14. You believe that a brand of electric fuses burns out from unpredictable power surges at the rate
of 1 every 6 months. You buy a packet that contains 7 fuses, and install them in a circuit one at a
time, replacing each fuse after it burns out. What is the probability that you will run out of these
fuses within 2 years time?
The idea is that we have a waiting time for a Poisson process, so this should follow a gamma
distribution. We are interested in all 7 fuses burning out, and a two-year span corresponds to 4 sixmonth periods. So as derived in class the probability that all burn out by period 4 is Gamma(7,1).
(Here 1 = 1.)
4

R 4 x6 ex
1
6 ex 6x5 ex 30x4 ex 120x3 ex 360x2 ex 640xex 640ex =
dx
=

x

7
720
0 1 6!
0
0.110674.

You might also like