

Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Material Type: Exam; Professor: Smirni; Class: Simulation; Subject: Computer Science; University: William and Mary; Term: Fall 2008;
Typology: Exams
1 / 2
This page cannot be seen from the preview
Don't miss anything!
This exam is due at the beginning of class (2:00pm) on Tuesday, November 4 and will not be accepted late without prior approval. Graduate students must work all three problems; undergraduate students must work problem 1 and either of the remaining two problems.
You may not seek or receive assistance on this test from any person other than me. Failure to observe this rule will be treated as an honor code violation.
Problem 1: [20 points] Mr. K is the manager of a new Snowball Express sno-cone stand which will be opening for business. The stand is operated by a single attendant who is behind a window. People come up to the window, tell the attendant which flavor sno-cone they want, wait for the attendant to fill their order, then pay and leave. When there is more than one customer at the window, a line forms so that customers are served in the order that they arrived. People always buy one and only one sno-cone at $3 each. The cost of materials to make sno-cones is negligible, so Mr. K considers all of the money taken in to be profit, except that the attendant has to be paid from the money received. After interviewing several applicants for the attendant’s position, Mr. K has narrowed his choice to two people: Fran and Bill. Fran can complete a transaction (take the order, make the sno-cone, take payment and make change) in 20 seconds on the average. Bill can perform the same job in 30 seconds, on the average. Fran is faster than Bill but demands $12 per hour, whereas Bill will work for $6 per hour. Market research has revealed that, on the average, Mr. K can expect one customer per minute to come up to the window. The research also shows that if a customer comes to the window and there are already N people in line, then the customer will promptly turn about in a huff and storm off to Ben and Jerry’s. Studies of real sno-cone stands have further shown that, even though the average customer arrival rate is one per minute, usually the amount of time between customers is less than that. It’s the occasional lull in business that makes the average time between customers longer than that which is usually observed. The same is true of the time it takes to make sno-cones. Thus, assume that the time between customer arrivals and the time it takes a person to make a sno-cone are exponentially distributed random values. (a) For both Fran and Bill, and based on 1 000 000 served customers, construct a table of the estimated steady-state probability that a customer will go to Ben and Jerry’s when the number of people in line N is 1, 2, 3, 4, and 5. (b) Construct similar tables if the time to serve a customer for Fran is changed to be Uniform(5, 35) and the time to serve a customer for Bill changed to be Uniform(15,45). (c) Comment on how the probability of rejection (i.e., the probability to go to Ben and Jerry’s) depends on the service process. (d) Assuming exponentially distributed service times for both Fran and Bill, if N is equal to 3, which person is the more cost-effective to hire: Fran or Bill? For this part of the problem provide also analytical solutions. (e) Discuss what you did to convince yourself that your results are correct.
Problem 2: [20 points] According to the ANSI C standard, the type unsigned long is guaranteed to support integers in the range 0, 1 ,... , 232 − 1.
Problem 3: [20 points] A test is compiled by selecting 12 different questions, at random and without replacement, from a well-publicized list of 120 questions. After studying this list you are able to classify all 120 questions into two classes, I and II. Class I questions are those about which you feel confident; the remaining questions define class II. Assume that your grade probability, conditioned on the class of the problems, is
A B C D F class I 0.6 0.3 0.1 0.0 0. class II 0.0 0.1 0.4 0.4 0.
Each test question is graded on an A = 4, B = 3, C = 2, D = 1, F = 0 scale and a score of 36 or better is required to pass the test. (a) If there are 90 class I questions in the list, use Monte Carlo simulation and 100 000 replications to generate a discrete-data histogram of scores. (b) Based on this histogram what is the probability that you will pass the test?
Evgenia Smirni, 10-20-