This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: STATS 320/725 Applied Stochastic Modelling Assignment 4 Due 4 pm, Thursday, 5 June, 2008 Note: This is the compulsory computing assignment. 1 In this question we will again be considering the single server queue with Poisson arrivals at rate 3 per unit time and service rate 4 per unit time. This time we will do our simulations using the method of batched means. There is a sample program for a simulation of an M/M/ 1 queue using the method of batched means attached to this assignment (see also Cecil). a Modify the script for batched means simulation to produce a batched means estimate of the average number in the system, L , in steady state in an M  G  1 queue with arrival rate 3 per unit time. The service time distribution should be changed so that with probability 2/3 a service time is equal to 1/8, and with probability 1/3 it is equal to 1/2. Set up the script so the simulation will have an initial transient period of 100 time units and then carry out 100 subruns of 200 time units each. Attach a copy of your R or Matlab script to the assignment. b Let X i denote the average queue length in the i th subrun. i Give the output that you observed from your simulation run, i.e. print X 1 ,X 2 ,...,X 100 . ii Give the point estimate, ¯ X , of L . Assuming the subruns are indepen dent, give a 95% confidence interval for L . iii Use your estimates for L to give a point estimate and 95% confidence interval for the time that a customer spends in the queue waiting for service (not including the service time). iv Give estimates, ˆ γ ( k ) of Cov ( X i ,X i + k ) for k = 1 , 2 , 3 , 4 , 5 , 6, and use acf to plot the autocorrelations. Is there any evidence that the sub runs are not independent? If there is, modify the 95% confidenceindependent?...
View
Full
Document
This note was uploaded on 09/24/2009 for the course STATS 341 taught by Professor Andrew during the Spring '09 term at Auckland.
 Spring '09
 ANDREW

Click to edit the document details