Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Short Test 4 – Probability Theory course 2023 – Solutions

Friday May 12th, 16:30–17:15

Full name: Student number: TA name:


• Write your full name and student number on both sheets provided, together with the name of your TA.

• This pair of sheets is the only one that you need to hand in to your TA at the end of the test. Please write your answers
starting with the back of this sheet. If you run out of space, ask the TA for an extra sheet and do not forget to also write there
your name and student number.

• No material of any kind is allowed during the short test. In particular, no book and no notes.

• Scrap paper is provided upon request. Any solution written on scrap paper will not be graded, make sure to copy it on this sheet.

• You may use a simple calculator for this test, but this is not necessary. In fact, you do not have to work out powers, fractions,
binomial coefficients, products, etc. Phones are not allowed (and so is their built-in calculator).

• Use proper notation and terminology, justify all your steps, explain your answers clearly, and do not just write the numerical
answer.

• You have 45 minutes available for the test. You cannot leave the short test before 17:00. If you want to leave after 17:00 but
before the end of the test, raise your hand, and wait for the TA to collect your sheet.

• Your short test grade is given by 1+ the number of points (thus 10 is the maximum), rounded to the 1st decimal.

Exercise 1 [3.5 points]


Consider two continuous random variables X and Y with joint probability density
{︄
10xy 2 if 0 ≤ y ≤ 1 and 0 ≤ x ≤ y,
fX,Y (x, y) =
0 otherwise.

(a) [1 points] Check that fX,Y (x, y) is a joint probability density.


(b) [2 points] Find the two marginals probability densities fX (x) and fY (y).
(c) [0.5 points] Are X and Y independent? Argue why in either case.

Solution
(a) The given two-variate function is always nonnegative and
+∞ ∫︂ +∞ 1 ∫︂ y 1 ]︃y 1 ]︃1
x2 y5
∫︂ ∫︂ ∫︂ [︃ ∫︂ [︃
2 2 4
fX,Y (x, y) dx dy = 10xy dx dy = 10y dy = 5 y dy = 5 = 1.
−∞ −∞ 0 0 0 2 0 0 5 0

∫︂ +∞
(b) The marginal density is calculated using the formula fX (x) = fX,Y (x, y) dy.
−∞

If x > 1 or x < 0, fX,Y (x, y) = 0 and thus trivially fX (x) = 0. If 0 ≤ x ≤ 1, then the integrand is nonzero
only when y ≥ x and hence
1 1 ]︃1
y3
∫︂ ∫︂ [︃
2 2 10
fX (x) = 10xy dy = 10x y dy = 10x = (x − x4 ).
x x 3 x 3

Therefore, {︄
10
3 (x − x4 ) if 0 ≤ x ≤ 1,
fX (x) =
0 otherwise.
∫︂ +∞
We can find the marginal density fY (x) using the formula fY (x) = fX,Y (x, y) dx.
−∞

If y > 1 or y < 0, fX,Y (x, y) = 0 and thus trivially fY (y) = 0. If 0 ≤ y ≤ 1, then the integrand is nonzero
only when x ≤ y and hence
∫︂ y ∫︂ y
2 2
]︁y
x dx = 10y 2 x2 /2 0 = 10y 2 (y 2 /2 − 02 /2) = 5y 4 .
[︁
fY (y) = 10xy dx = 10y
0 0

Therefore, {︄
5y 4 if 0 ≤ y ≤ 1,
fY (y) =
0 otherwise.

(c) X and Y are not independent because, using the marginals derived in (b), one can check that the
equality fX,Y (x, y) = fX (x) · fY (y) does not hold for all (x, y) ∈ R2 , e.g., for (x, y) = (1, 1).

Exercise 2 [5.5 points]


Alex and Barbara are running as a team in a 10-mile relay race. The time X Alex takes to finish (in hours)
is uniformly distributed, specifically X ∼ U([0, 2]), while the time Y Barbara takes to finish her mile (in
hours) is exponentially distributed, specifically Y ∼ Exp(λ).
(a) [0.5 points] Write the p.d.f.’s fX and fY of the r.v.’s X and Y . Note: You do not need to derive them.
(b) [0.5 points] Calculate the mean and the variance of both X and Y .
Note: You do not need to derive them from the densities if you remember the formulas.
Barbara starts immediately after Alex finishes his mile, and their performances (and hence their times) are
independent. Consider the continuous random variable Z = X + Y that describes the total time they take
to finish the relay race.
(c) [2.25 points] What is the support of Z? Find the distribution of Z = X + Y by explicitly deriving its
density fZ (z). Hint: it is helpful to distinguish three cases: z < 0, 0 ≤ z ≤ 2 and z > 2
(d) [0.5 points] Calculate the joint probability density fX,Y (x, y).
(e) [1.25 points] Calculate the probability that Alex is slower, that is P(X < Y ).
Let U and V be two independent normal r.v.’s with the same mean and variance as X and Y , respectively.
(f) [0.5 points] Using your answers from (b), conclude what is the distribution of the r.v. W = U + V .
Note: You can fully answer this question even if you didn’t finish questions (c), (d), and (e).

Solution
(a) The p.d.f. of X is fX (x) = 1
2 for x ∈ [0, 2] and 0 otherwise. The p.d.f. of Y is fY (y) = λe−λy for y ≥ 0
and 0 otherwise.
(b) Using the standard formulas for mean and variance of uniform and exponential distributions, we get
EX = 1, EY = λ1 , VarX = 124
= 13 , VarY = λ12 .
(c) The support of X is [0, 2] and that of Y is [0, +∞), therefore the support of Z = X + Y is [0, +∞).
We derive fZ using the following convolution formula:
∫︂ +∞
fZ (z) = fX ∗ fY (z) = fX (x)fY (z − x) dx.
−∞
If z < 0, then clearly fZ (z) = 0 since it is impossible for the r.v. Z to take negative values, being the sum
of two nonnegative r.v.’s.
For fixed z, we need to check for which values the density fY (z − x) is nonzero. We distinguish two cases:
• If z ∈ [0, 2], then z − x ≥ 0 only if x ≤ z, so we calculate
∫︂ z
1 1 [︂ ]︂z 1
fZ (z) = · λe−λ(z−x) dx = e−λz eλx = (1 − e−λz ).
0 2 2 0 2

• If z ∈ (2, +∞), then z − x ≥ 0 for every x, since fX takes nonzero values only for x ∈ [0, 2]. In this
case, we can calculate
∫︂ 2
1 1 [︂ ]︂2 1
fZ (z) = · λe−λ(z−x) dx = e−λz eλx = e−λz (e2λ − 1).
0 2 2 0 2

The resulting density for the r.v. Z then is

if z < 0,

⎨0

fZ (z) = 12 (1 − e−λz ) if z ∈ [0, 2],
if z ∈ (2, +∞).

⎩ 1 −λz 2λ
2e (e − 1)

(d) Since the two random variables are independent, the joint density is simply the product of the two
marginal densities, and hence
{︄
1
λe−λy if y ≥ 0 and 0 ≤ x ≤ 2,
fX,Y (x, y) = 2
0 otherwise.

(e) The required probability can be obtained by integrating the joint density over the set B = {(x, y) ∈
R2 : x < y}, hence

e−λ sinh(λ) 1 − e−2λ


∫︂ ∫︂ 2 (︃∫︂ +∞ )︃ ∫︂ +∞
1 −λy 1 −λx
P(X < Y ) = fX,Y (x, y) dx dy = λe dy dx = e dx = = .
B 0 x 2 0 2 λ 2λ

(f) In view of (b), we get U ∼ N (1, 1/3) and V ∼ N (λ−1 , λ−2 ). The sum of two independent normal r.v.’s
is again a normal r.v. with mean µ = 1 + λ−1 and variance σ 2 = 1/3 + λ−2 .

You might also like