Massachusetts Institute of Technology: (Final Exam - Spring 2009)

Massachusetts Institute of Technology
Department of Electrical Engineering & Computer Science

6.041/6.431: Probabilistic Systems Analysis
(Final Exam | Spring 2009)
Problem 1: True or False (2pts. each, 18 pts. total)
No partial credit will be given for individual questions in this part of the quiz.
a. Let {Xn } be a sequence of i.i.d random variables taking values in the interval [0, 0.5]. Consider
the following statements:
(A) If E[Xn2 ] converges to 0 as n → ∞ then Xn converges to 0 in probability.
(B) If all Xn have E[Xn ] = 0.2 and var (Xn ) converges to 0 as n → ∞ then Xn converges to 0.2
in probability.
(C) The sequence of random variables Zn , defined by Zn = X1 · X2 · · · Xn , converges to 0 in
probability as n → ∞.
Which of these statements are always true? Write True or False in each of the boxes below.
A: B: C:
b. Let Xi (i = 1, 2, . . . ) be i.i.d. random variables with mean 0 and variance 2; Yi (i = 1, 2, . . . ) be

i.i.d. random variables with mean 2. Assume that all variables Xi , Yj are independent. Consider
the following statements:
X1 +···+Xn
(A) n converges to 0 in probability as n → ∞.
X12 +···+Xn2
(B) n converges to 2 in probability as n → ∞.
X1 Y1 +···+Xn Yn
(C) n converges to 0 in probability as n → ∞.
A: B: C:
c. We have i.i.d. random variables X1 . . . Xn with an unknown distribution, and with µ = E[Xi ].
We define Mn = (X1 + . . . + Xn )/n. Consider the following statements:
(A) Mn is a maximum-likelihood estimator for µ, irrespective of the distribution of the Xi ’s.

(B) Mn is a consistent estimator for µ, irrespective of the distribution of the Xi ’s.
(C) Mn is an asymptotically unbiased estimator for µ, irrespective of the distribution of the
Xi ’s.
A: B: C:
1
Problem 2: Multiple Choice (4 pts. each, 24 pts. total)
Clearly circle the appropriate choice. No partial credit will be given for individual questions in this
part of the quiz.
a. Earthquakes in Sumatra occur according to a Poisson process of rate λ = 2/year. Conditioned

on the event that exactly two earthquakes take place in a year, what is the probability that both
earthquakes occur in the first three months of the year? (for simplicity, assume all months have
30 days, and each year has 12 months, i.e., 360 days).
(i) 1/12
(ii) 1/16
(iii) 64/225
(iv) 4e−4
(v) There is not enough information to determine the required probability.
(vi) None of the above.
b. Consider a continuous-time Markov chain with three states i ∈ {1, 2, 3}, with dwelling time in
each visit to state i being an exponential random variable with parameter νi = i, and transition
probabilities pij defined by the graph
What is the long-term expected fraction of time spent in state 2?
(i) 1/2
(ii) 1/4
(iii) 2/5
(iv) 3/7
(v) None of the above.
2
c. Consider the following Markov chain:
Starting in state 3, what is the steady-state probability of being in state 1?
(i) 1/3
(ii) 1/4
(iii) 1
(iv) 0
d. Random variables X and Y are such that the pair (X, Y ) is uniformly distributed over the
trapezoid A with corners (0, 0), (1, 2), (3, 2), and (4, 0) shown in Fig. 1:
X
1 3 Y
4
Figure 1: fX,Y (x, y) is constant over the shaded area, zero otherwise.
i.e. �
c, (x, y) ∈ A
fX,Y (x, y) =
0, else .
We observe Y and use it to estimate X. Let X̂ be the least mean squared error estimator of X
given Y . What is the value of var(X̂ − X|Y = 1)?
(i) 1/6
(ii) 3/2
(iii) 1/3
(iv) The information is not sufficient to compute this value.
3
e. X1 . . . Xn are i.i.d. normal random variables with mean value µ and variance v. Both µ and v
are unknown. We define Mn = (X1 + . . . + Xn )/n and
n
1 �
Vn = (Xi − Mn )2
n−1
i=1
We also define Φ(x) to be the CDF for the standard normal distribution, and Ψn−1 (x) to be the
CDF for the t-distribution with n − 1 degrees of freedom. Which of the following choices gives
an exact 99% confidence interval for µ for all n > 1?
� �
(i) [Mn − δ Vnn , Mn + δ Vnn ] where δ is chosen to give Φ(δ) = 0.99.
� �
(ii) [Mn − δ Vnn , Mn + δ Vnn ] where δ is chosen to give Φ(δ) = 0.995.
� �
(iii) [Mn − δ n , Mn + δ Vnn ] where δ is chosen to give Ψn−1 (δ) = 0.99.
Vn
� �
(iv) [Mn − δ Vnn , Mn + δ Vnn ] where δ is chosen to give Ψn−1 (δ) = 0.995.
f. We have i.i.d. random variables X1 , X2 which have an exponential distribution with unknown
parameter θ. Under hypothesis H0 , θ = 1. Under hypothesis H1 , θ = 2. Under a likelihood-ratio
test, the rejection region takes which of the following forms?
(i) R = {(x1 , x2 ) : x1 + x2 > ξ} for some value ξ.

(ii) R = {(x1 , x2 ) : x1 + x2 < ξ} for some value ξ.
(iii) R = {(x1 , x2 ) : ex1 + ex2 > ξ} for some value ξ.
(iv) R = {(x1 , x2 ) : ex1 + ex2 < ξ} for some value ξ.
4
Problem 3 (12 pts. total)
Aliens of two races (blue and green) are arriving on Earth independently according to Poisson process
distributions with parameters λb and λa respectively. The Alien Arrival Registration Service Authority
(AARSA) will begin registering alien arrivals soon.
Let T1 denote the time AARSA will function until it registers its first alien. Let G be the event
that the first alien to be registered is a green one. Let T2 be the time AARSA will function until at
least one alien of both races is registered.
(a) (4 points.) Express µ1 = E[T1 ] in terms of λg and λb . Show your work.
(b) (4 points.) Express p = P(G) in terms of λg and λb . Show your work.
5
(c) (4 points.) Express µ2 = E[T2 ] in terms of λg and λb . Show your work.
6
Researcher Jill is interested in studying employment in technology firms in Dilicon Valley. She denotes
by Xi the number of employees in technology firm i and assumes that Xi are independent and identically
distributed with mean p. To estimate p, Jill randomly interviews n technology firms and observes the
number of employees in these firms.
(a) (6 points.) Jill uses

X1 + · · · + Xn
Mn =
n
as an estimator for p. Find the limit of P(Mn ≤ x) as n → ∞ for x < p. Find the limit of
P(Mn ≤ x) as n → ∞ for x > p. Show your work.
(b) (6 points.) Find the smallest n, the number of technology firms Jill must sample, for which the
Chebyshev inequality yields a guarantee
P(|Mn − p| ≥ 0.5) ≤ 0.05.
Assume that var (Xi ) = v for some constant v. State your solution as a function of v. Show
your work.
7
(c) (6 points.) Assume now that the researcher samples n = 5000 firms. Find an approximate value
for the probability
P(|M5000 − p| ≥ 0.5)
using the Central Limit Theorem. Assume again that var (Xi ) = v for some constant v. Give
your answer in terms of v, and the standard normal CDF Φ. Show your work.
8
The RandomView window factory produces window panes. After manufacturing, 1000 panes were
loaded onto a truck. The weight Wi of the i-th pane (in pounds) on the truck is modeled as a random
variable, with the assumption that the Wi ’s are independent and identically distributed.
(a) (6 points.) Assume that the measured weight of the load on the truck was 2340 pounds, and
that var (Wi ) ≤ 4. Find an approximate 95 percent confidence interval for µ = E[Wi ], using the
Central Limit Theorem (you may use the standard normal table which was handed out with this
quiz). Show your work.
(b) (6 points.) Now assume instead that the random variables Wi are i.i.d, with an exponential
distribution with parameter θ > 0, i.e., a distribution with PDF
fW (w; θ) = θe−θw
What is the maximum likelihood estimate of θ, given that the truckload has weight 2340 pounds?
Show your work.
9
In Alice’s Wonderland, there are six different seasons: Fall (F), Winter (W), Spring (Sp), Summer (Su),
Bitter Cold (B), and Golden Sunshine (G). The seasons do not follow any particular order, instead, at
the beginning of each day the Head Wizard assigns the season for the day, according to the following
Markov chain model:
Thus, for example, if it is Fall one day then there is 1/6 probability that it will be Winter the next
day (note that it is possible to have the same season again the next day).
(a) (4 points.) For each state in the above chain, identify whether it is recurrent or transient. Show
your work.
(b) (4 points.) If it is Fall on Monday, what is the probability that it will be Summer on Thursday
of the same week? Show your work.
10
(c) (4 points.) If it is Spring today, will the chain converge to steady-state probabilities? If so,
compute the steady-state probability for each state. If not, explain why these probabilities do
not exist. Show your work.
(d) (5 points.) If it is Fall today, what is the probability that Bitter Cold will never arrive in the
future? Show your work.
11
(e) (5 points.) If it is Fall today, what is the expected number of days till either Summer or Golden
Sunshine arrives for the first time? Show your work.
12
A newscast covering the final baseball game between Sed Rox and Y Nakee becomes noisy at the
crucial moment when the viewers are informed whether Y Nakee won the game.
Let a be the parameter describing the actual outcome: a = 1 if Y Nakee won, a = −1 otherwise.
There were n viewers listening to the telecast. Let Yi be the information received by viewer i (1 ≤ i ≤
n). Under the noisy telecast, Yi = a with probability p, and Yi = −a with probability 1 − p. Assume
that the random variables Yi are independent of each other.
The viewers as a group come up with a joint estimator
� �n
1 if i=1 Yi ≥ 0,
Zn =
−1 otherwise.
(a) (6 points.)
Find limn→∞ P(Zn = a) assuming that p > 0.5 and a = 1. Show your work.
(b) (6 points.) Find limn→∞ P(Zn = a), assuming that p = 0.5 and a = 1. Show your work.
13
MIT OpenCourseWare
http://ocw.mit.edu
6.041SC Probabilistic Systems Analysis and Applied Probability

Fall 2013
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Massachusetts Institute of Technology: (Final Exam - Spring 2009)

Uploaded by

Copyright:

Available Formats

Massachusetts Institute of Technology: (Final Exam - Spring 2009)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Massachusetts Institute of Technology: (Final Exam - Spring 2009)

Uploaded by

Copyright:

Available Formats

Massachusetts Institute of Technology

Department of Electrical Engineering & Computer Science

Problem 1: True or False (2pts. each, 18 pts. total)

b. Let Xi (i = 1, 2, . . . ) be i.i.d. random variables with mean 0 and variance 2; Yi (i = 1, 2, . . . ) be

(A) Mn is a maximum-likelihood estimator for µ, irrespective of the distribution of the Xi ’s.

Problem 2: Multiple Choice (4 pts. each, 24 pts. total)

a. Earthquakes in Sumatra occur according to a Poisson process of rate λ = 2/year. Conditioned

What is the long-term expected fraction of time spent in state 2?

c. Consider the following Markov chain:

Starting in state 3, what is the steady-state probability of being in state 1?

(i) R = {(x1 , x2 ) : x1 + x2 > ξ} for some value ξ.

Problem 3 (12 pts. total)

(a) (4 points.) Express µ1 = E[T1 ] in terms of λg and λb . Show your work.

(b) (4 points.) Express p = P(G) in terms of λg and λb . Show your work.

(c) (4 points.) Express µ2 = E[T2 ] in terms of λg and λb . Show your work.

Problem 4 (18 pts. total)

(a) (6 points.) Jill uses

P(|Mn − p| ≥ 0.5) ≤ 0.05.

Problem 5 (12 pts. total)

Problem 6 (21 pts. total)

Problem 7 (12 pts. total)

6.041SC Probabilistic Systems Analysis and Applied Probability

You might also like