6.

041 Fall 2009 Final Exam

Name: Recitation Instructor: TA:

Question 1 2 (a) 2 (b) 2 (c) 2 (d) 3 (a) 3 (b) 3 (c) 3 (d) 3 (e)

Score

Out of 5 5 5 5 5 5 5 5 5

Question 4 (a) 4 (b) 4 (c) 4 (d) 4 (e) 4 (f ) 5 (a) 5 (b) 5 (c) 5 (d) 5 (e) Your Grade

Score

Out of 5 5 5 5 5 5 5 5 5 5 5 100

This exam has 5 problems, worth a total of 100 points. When giving a formula for a PDF, make sure to specify the range over which the formula holds. Write your solutions in this quiz booklet, only solutions in this quiz booklet will be graded. Be neat! You will not get credit if we cant read it. You are allowed three two-sided, handwritten, formula sheets plus a calculator. You may give an answer in the form of an arithmetic expression (sums, products, ratios, factorials) of numbers that could be evaluated using a calculator. Expressions like 8 or 3 5 k k=0 (1/2) are also ne. The last page of this nal contains a standard normal table.

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
Problem 2. (20 points)
A pair of jointly continuous random variables, X and Y , have a joint probability density function given by
c, in the shaded region of Fig. 1 fX,Y (x, y ) = 0, elsewhere.

Massachusetts Institute of Technology

y
1 1/2

1 /2

Figure 1: The shaded region is the domain in which fX,Y (x, y ) = c. (a) (5 points) Find c. (b) (5 points) Find the marginal PDFs of X and Y , i.e., fX (x) and fY (y ). (c) (5 points) Find E[X | Y = 1/4] and Var[X | Y = 1/4], that is, the conditional mean and conditional variance of X given that Y = 1/4. (d) (5 points) Find the conditional PDF for X given that Y = 3/4, i.e., fX |Y (x | 3/4). Problem 3. (25 points)
Consider a Markov chain Xn whose one-step transition probabilities are shown in the gure.

1 4 1 3

1
1 4 1 3

1 6

1 3
1 3
(a) (5 points) What are the recurrent states?

4
1 3

Page 3 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(b) (5 points) Find P(X2 = 4 | X0 = 2). (c) (5 points) Suppose that you are given the values of rij (n) = P(Xn = j | X0 = i). Give a formula for r11 (n + 1) in terms of the rij (n). (d) (5 points) Find the steady-state probabilities j = limn P(Xn = j | X0 = i), or explain why they do not exist. (e) (5 points) What is the probability of eventually visiting state 4, given that the initial state is X0 = 1? Problem 4. (30 points) Al, Bonnie, and Clyde run laps around a track, with the duration of each lap (in hours) being exponentially distributed with parameters A = 21, B = 23, and C = 24, respectively. Assume that all lap durations are independent. At the completion of each lap, a runner drinks either one or two cups of water, with probabilities 1/3 and 2/3, respectively, independent of everything else, including how much water was consumed after previous laps. (The time spent drinking is negligible, assumed zero.) (a) (5 points) Write down the PMF of the total number of completed laps over the rst hour. (b) (5 points) What is the expected number of cups of water to be consumed by the three runners, in total, over the rst hour. (c) (5 points) Al has amazing endurance and completed 72 laps. Find a good approximation for the probability that he drank at least 130 cups. (You do not have to use 1/2-corrections.) (d) (5 points) What is the probability that Al nishes his rst lap before any of the others? (e) (5 points) Suppose that the runners have been running for a very long time when you arrive at the track. What is the distribution of the duration of Als current lap? (This includes the duration of that lap both before and after the time of your arrival.) (f) (5 points) Suppose that the runners have been running for 1/4 hours. What is the distribution of the time Al spends on his second lap, given that he is on his second lap? Problem 5. (25 points)
A pulse of light has energy X that is a second-order Erlang random variable with parameter , i.e., its PDF is
2 xex , for x 0, fX (x) = (1) 0, otherwise. This pulse illuminates an ideal photon-counting detector whose output N is a Poisson-distributed random variable with mean x when X = x, i.e., its conditional PMF is n x x e , for n = 0, 1, 2, . . . , n! pN |X (n | x) = 0, otherwise. (a) (5 points) Find E[N ] and Var[N ], the unconditional mean and variance of N (b) (5 points) Find pN (n), the unconditional PMF of N . lin (N ), the linear least-squares estimator of X based on an observation of N . (c) (5 points) Find X MAP (N ), the MAP estimator of X based on an observation of N . (d) (5 points) Find X

Massachusetts Institute of Technology

Page 4 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(e) (5 points) Instead of the prior distribution in Eq. (1), we are now told that
P(X = 2) = 33 /35, P(X = 3) = 23 /35.

Massachusetts Institute of Technology

Given the observation N = 3, and in order to minimize the probability of error, which one of the two hypotheses X = 2 and X = 3 should be chosen? Useful integral and facts: y k ey dy =
0

k! k+1

for > 0 and k = 0, 1, 2, . . . (recall that 0!=1)

The second-order Erlang random variable satises: E[X ] = 2/, Var(X ) = 2/2 .

Each question is repeated in the following pages. Please write your answer on
the appropriate page.

Page 5 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
Problem 2. (20 points)
A pair of jointly continuous random variables, X and Y , have a joint probability density function given by
c, in the shaded region of Fig. 1 fX,Y (x, y ) = 0, elsewhere.

Massachusetts Institute of Technology

y
1 1/2

1 /2

Figure 2: The shaded region is the domain in which fX,Y (x, y ) = c. (a) (5 points) Find c.

(b) (5 points) Find the marginal PDFs of X and Y , i.e., fX (x) and fY (y ).

Page 6 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(c) (5 points) Find E[X | Y = 1/4] and Var[X | Y = 1/4], that is, the conditional mean and conditional variance of X given that Y = 1/4.

Massachusetts Institute of Technology

(d) (5 points) Find the conditional PDF for X given that Y = 3/4, i.e., fX |Y (x | 3/4).

Page 7 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
Problem 3. (25 points)
Consider a Markov chain Xn whose one-step transition probabilities are shown in the gure.

Massachusetts Institute of Technology

1 4 1 3

1
1 4 1 3

1 6

1 3
1 3
(a) (5 points) What are the recurrent states?

4
1 3

(b) (5 points) Find P(X2 = 4 | X0 = 2).

Page 8 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(c) (5 points) Suppose that you are given the values of rij (n) = P(Xn = j | X0 = i). Give a formula for r11 (n + 1) in terms of the rij (n).

Massachusetts Institute of Technology

(d) (5 points) Find the steady-state probabilities j = limn P(Xn = j | X0 = i), or explain why they do not exist.

Page 9 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(e) (5 points) What is the probability of eventually visiting state 4, given that the initial state is X0 = 1?

Massachusetts Institute of Technology

Page 10 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
Problem 4. (30 points) Al, Bonnie, and Clyde run laps around a track, with the duration of each lap (in hours) being exponentially distributed with parameters A = 21, B = 23, and C = 24, respectively. Assume that all lap durations are independent. At the completion of each lap, a runner drinks either one or two cups of water, with probabilities 1/3 and 2/3, respectively, independent of everything else, including how much water was consumed after previous laps. (The time spent drinking is negligible, assumed zero.) (a) (5 points) Write down the PMF of the total number of completed laps over the rst hour.

Massachusetts Institute of Technology

(b) (5 points) What is the expected number of cups of water to be consumed by the three runners, in total, over the rst hour.

Page 11 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(c) (5 points) Al has amazing endurance and completed 72 laps. Find a good approximation for the probability that he drank at least 130 cups. (You do not have to use 1/2-corrections.)

Massachusetts Institute of Technology

(d) (5 points) What is the probability that Al nishes his rst lap before any of the others?

Page 12 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(e) (5 points) Suppose that the runners have been running for a very long time when you arrive at the track. What is the distribution of the duration of Als current lap? (This includes the duration of that lap both before and after the time of your arrival.)

Massachusetts Institute of Technology

(f) (5 points) Suppose that the runners have been running for 1/4 hours. What is the distribution of the time Al spends on his second lap, given that he is on his second lap?

Page 13 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
Problem 5. (25 points)
A pulse of light has energy X that is a second-order Erlang random variable with parameter , i.e., its PDF is
2 xex , for x 0, fX (x) = (1) 0, otherwise. This pulse illuminates an ideal photon-counting detector whose output N is a Poisson-distributed random variable with mean x when X = x, i.e., its conditional PMF is n x x e , for n = 0, 1, 2, . . . , n! pN |X (n | x) = 0, otherwise. Useful integral and facts: y k ey dy =
0

Massachusetts Institute of Technology

k! k+1

for > 0 and k = 0, 1, 2, . . . (recall that 0!=1)

The second-order Erlang random variable satises: E[X ] = 2/, Var(X ) = 2/2 .

(a) (5 points) Find E[N ] and Var[N ], the unconditional mean and variance of N

Page 14 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
(b) (5 points) Find pN (n), the unconditional PMF of N .

Massachusetts Institute of Technology

lin (N ), the linear least-squares estimator of X based on an observation of N . (c) (5 points) Find X

Page 15 of 16

Department of Electrical Engineering & Computer Science 6.041/6.431: Probabilistic Systems Analysis (Fall 2009)
MAP (N ), the MAP estimator of X based on an observation of N . (d) (5 points) Find X

Massachusetts Institute of Technology

(e) (5 points) Instead of the prior distribution in Eq. (1), we are now told that P(X = 2) = 33 /35, P(X = 3) = 23 /35.

Given the observation N = 3, and in order to minimize the probability of error, which one of the two hypotheses X = 2 and X = 3 should be chosen?

Page 16 of 16

MIT OpenCourseWare http://ocw.mit.edu

6.041 / 6.431 Probabilistic Systems Analysis and Applied Probability

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