Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
MET270 Introduction to Statistics and Probability 2, autumn 2010. Exercise set 8 Textbook (Ghahramani): Section 12.3: Exercise 1, 3, 7 Section 12, review problems: Exercise 1, 3, 4 Exercise 1: Suppose that the queue to be served at Idsø (a famous butcher in Stavanger) on a Saturday is a M/M/4 queue with arrival rate λ = 1.5 per minute and expected service time of 1/γ = 2 minutes. When you arrive four persons are being served and 10 are waiting in line. Suppose now that the chance of rain on a day does not only depend on the weather the previous day, but also on the weather two days ago. Then we can not use the simple Markov model above, but we can actually still model the situation as a Markov chain by augmenting the state space to contain the weather sequence the two last days (see also example 12.11 in Ghahramani). Let state 0: ss (sun-sun), state 1: rs (rain-sun), state 2: sr (sun-rain), state 3: rr (rain-rain). Let {Xn : n = 0, 1, 2, . . .} be a stochastic process with the four states above as state space. Suppose the chance of rain is 0.3 if it was not raining the two previous days, 0.5 if it was not raining the day before but was raining two days before, 0.6 if it was raining the day before but not two days before and 0.8 if it was raining the two previous days. b) What is the expected waiting time until you are starting to get served? What is the expected waiting time until you are finished? c) Explain why {Xn : n = 0, 1, 2, . . .} is matrix ⎛ 0.7 ⎜ ⎜ 0.5 P =⎜ ⎜ 0 ⎝ 0 c) What is the probability that you are finished before the person in front of you in the queue? What is the probability that the person behind you are finished before you? It is given that: a) Explain why the queue is stable. Would the queue still be stable if one of the four persons serving customers had to leave work? ⎛ Suppose that the chance of rain tomorrow only depends on whether it is raining today and not on past weather conditions. Let the states be labelled, 0: no rain, 1: rain. Suppose that: P (rain tomorrow|rain today) = 0.75 P (rain tomorrow|not rain today) = 0.35 ⎞ 0 0.3 0 ⎟ 0 0.5 0 ⎟ ⎟ 0.4 0 0.6 ⎟ ⎠ 0.2 0 0.8 0.49 0.12 0.20 0.12 0.10 0.16 ⎜ ⎜ 0.35 P2 = ⎜ ⎜ 0.20 ⎝ Exercise 2: a Markov chain with transition probability 0.21 0.15 0.20 0.10 ⎞ 0.18 ⎟ 0.30 ⎟ ⎟ 0.48 ⎟ ⎠ 0.64 d) It is raining today, but was not raining yesterday, what is the probability that it is not raining tomorrow? If it was not raining today and yesterday, what is the probability it will be raining both the two next days? If it was raining yesterday, but not today, what is the probability it will be raining tomorrow but not the day after? a) Find the transition probability matrix. Show that P2 = 0.51 0.49 0.35 0.65 Some answers: 12.3:1 Not Markov; 12.3:3 a) 3/8, b) 11/16; 12.3:7 Recurrent: {0, 3} and {2, 4}. Transient: {1}. b) If it is raining today, what is the probability that it is not raining tomorrow? If it is not raining today, what is the probability that it will be raining in two days? If it was raining yesterday, what is the probability that it will not be raining tomorrow? 12.review:1 0.692; 12.review:3 Rename 1 to 4, 3 to 1 and 4 to 3. 12.review:4 5/3. 1 a) no; b) 5.5 and 7.5; c) 3/8 and 3/8; 2 b) 0.25, 0.49 and 0.35; d) 0.4, 0.18, and 0.20.