Math 20: Discrete Probability Exam Solutions - Bernoulli Trials and Markov Chains, Exams of Probability and Statistics

Solutions to the final exam of a discrete probability course, covering topics such as bernoulli trials, expected values, and variance. It also includes a problem related to a markov chain model of a political election and a popular politician's re-election probability.

Typology: Exams

2012/2013

Uploaded on 02/20/2013

rangith
rangith 🇮🇳

4.7

(7)

46 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Math 20: Discrete Probability
Final Exam Solutions
December 3, 2000
1Consider a Bernoulli trials process with probability pfor success (and
probability q= 1pfor failure). Let Sndenote the total number of successes
in the first ntrials, and let An=Sn/n denote the average number of successes
in the first ntrials.
(a) Show that E(Sn) = np and E(An) = p.
(b) Show that V(Sn) = npq and V(An) = pq
n.
2In the current presidential election, 100,000,000 people voted, and Gore
came out with about 200,000 more votes. Assume that the voting is a
Bernoulli trials with probability pthat a given voter votes for Gore. If
p=1
2, estimate the probability that Gore’s total would be as high as it is
(i.e. greater than or equal to 50,100,000).
3More voting! A popular politician runs for Congress. If she has never
been elected, then the probability that she will be elected is 1
2(and so the
probability that she remains unelected is 1
2and she can run again next time,
in two years). If she has already been elected (and is currently in office) then
her probabiltiy of being re-elected is 9
10 ; the probability that she loses is 1
10 .
If she loses, then she will never be re-elected again, so she retires.
(a) Show how to think of this as a Markov chain. That is, write down the
states and the transition matrix. Explain why the Markov chain is an
absorbing one.
(b) If this is the first year that she runs for Congress, in how many years
should she expect to retire?
1
pf2

Partial preview of the text

Download Math 20: Discrete Probability Exam Solutions - Bernoulli Trials and Markov Chains and more Exams Probability and Statistics in PDF only on Docsity!

Math 20: Discrete Probability

Final Exam Solutions December 3, 2000

1 Consider a Bernoulli trials process with probability p for success (and probability q = 1−p for failure). Let Sn denote the total number of successes in the first n trials, and let An = Sn/n denote the average number of successes in the first n trials.

(a) Show that E(Sn) = np and E(An) = p.

(b) Show that V (Sn) = npq and V (An) = pq n.

2 In the current presidential election, 100, 000 , 000 people voted, and Gore came out with about 200, 000 more votes. Assume that the voting is a Bernoulli trials with probability p that a given voter votes for Gore. If p = 12 , estimate the probability that Gore’s total would be as high as it is (i.e. greater than or equal to 50, 100 , 000).

3 More voting! A popular politician runs for Congress. If she has never been elected, then the probability that she will be elected is 12 (and so the probability that she remains unelected is 12 and she can run again next time, in two years). If she has already been elected (and is currently in office) then her probabiltiy of being re-elected is 109 ; the probability that she loses is 101. If she loses, then she will never be re-elected again, so she retires.

(a) Show how to think of this as a Markov chain. That is, write down the states and the transition matrix. Explain why the Markov chain is an absorbing one.

(b) If this is the first year that she runs for Congress, in how many years should she expect to retire?

4 The following matrix is the transition matrix for an absorbing Markov chain. The first transient state is state S, the second is state T.

P =

     

2 5 0 0

1 5

1 5

1 5 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1

     

(a) If the chain starts at state S how many steps do you expect it will it take until the chain lands in an absorbing state?

(b) Again assuming that the chain starts in state S, find the likelihood of being absorbed in any given absorbing state.

(c) Suppose we start in state S with probability 13 and in state T with probability 23. Find the likelihood of being in any given state after two iterations.

5 Give short answers to the following questions.

(a) If you toss a fair coin n times (where n is HUGE), does the Law of Large Numbers tell you that the total number of heads will differ from n 2 by no more than 1000? (b) Let Sn be the number of heads in n tosses of a fair coin. Find

nlim→∞ P

( Sn <

n 2

n

) .

(c) Let Sn be the number of heads in n tosses of a fair coin. Find

nlim→∞ P

( Sn <

n 2

n

) .

(d) Is this a cool class or what?

6 You roll a fair die 600 times, so you expect five to come up 100 times. Find a number x so that the chances of there being between 100 − x and 100 + x is roughly 0.9.