Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

STOR 445 MIDTERM II

Thursday, Mar 31, 2022.

2:00-3:15am.

1. Attempt all problems. Show all work. Upload to gradescope by 3:30pm.

2. The problem weights are as given in the parentheses.

3. You are allowed one 8.5 by 11 sheet of paper with notes of your choice on both sides.

4. If you use Matlab or R programs state explicitly how you called them.

5. No communication about the exam is allowed with anybody except the proctor.

P1.  (30) Suppose Bob’s email inbox is empty at time 0.  An incoming email

message can be real or spam.  Bob gets 8 spam messages per hour, and 5 real messages per hour.  Suppose the arrival processes of the spam and real emails are independent Poisson processes.  Currently Bob does not have any spam filter, so both type of emails are put in the inbox.

(a)  (4) What is the distribution of the time when the first message arrives (it could be either real or spam) in the inbox?

(b)  (4) What is the probability that first message is spam?

(c)  (5) What is expected time until the arrival of the third real message?

(d)  (5) What is the variance of the total number of messages over four hours?

Bob gets annoyed by the spam messages and decides to add a spam filter. The spam filters blocks 80% of the spam messages and puts them in the spam folder.  Unfortunately, it also blocks 10% of ral messages and puts them in the same spam folder. The spam folder is empty at time 0.

(e)  (6) Let N(t) be the number of messages in the spam folder at time t. Is {N(t),t ≥ 0} a Poisson process? Why or why not? If it, what is its rate?

(f)  (6) If 20 spam messages arrive in the spam folder in the first four

hours, what is the probability that 3 of them are real messages?

P2.  (30) Mrs.  Smith suffers from a chronic pain condition that requires fre-

quent visits to the doctor.  She has an option of visiting the doctor in person or remotely. The in-person visits are more effective, but more in- convenient; while remote visits are convenient but not as effective. After a remote visit Mrs. Smith needs to revisit the doctor after about 25 days, while after in-person visit she needs to return in about 40 days. Suppose she opts for in-person visit about 30% of the visits, regardless of past his- tory.  Let X(t) be 1 if Mrs.  Smith’s most recent visit up to time t was in-person, and 2 if the most recent visit up to time t was remote. Assume the initial visit was in-person at time 0.

(a)  (10) State the assumptions needed to make {X(t),t ≥ 0} a CTMC. What is its state-space and the rate matrix?

(b)  (10) Suppose these assumptions in part (a) hold. Compute the prob- ability that the most recent visit of Mrs.  Smith up to time t = 50 was virtual.

(c)  (10) Suppose the virtual visit costs $100, while the in-person visit costs $140. What is the long run cost per day for Mrs. Smith?

P3.  (40) Consider a call center with 4 operators and no capacity to put callers

on hold.  Callers arrive according to a Poisson process with rate 50 per hour.  A caller that finds all operators busy is lost.  Those who find an available operator start their service immediately, which lasts for an ex- ponential amount of time with mean 5 minutes. Let X(t) be the number of busy operators at time t.

(a)  (10) Show that {X(t),t ≥ 0} is a birth and death process. What is its state space and what are the birth and death rates?

(b)  (10) Compute the expected number of busy operators in the long run.

(c)  (10) Suppose the operators are paid an hourly salary of 20 dollars whether they are busy or not.  Each caller who gets served pays 5 dollars in fees. Compute the long run net revenue generated by this system.

(d)  (10) The system manager is thinking of hiring one more operator. Do you think this is a good decision? Why or why not?