Download Examination Materials for Computer Simulation and Modelling in Software Development and more Exams Mathematical Modeling and Simulation in PDF only on Docsity!
CORK INSTITUTE OF TECHNOLOGY
INSTITIÚID TEICNEOLAÍOCHTA CHORCAÍ
Semester 2 Examinations 2011
Module Title: Computer Simulation and Modelling
Module Code: COMP 9005
School: Computing & Mathematics
Programme Title: Master of Science in Software Development
Programme Code: KSDEV_9_Y
External Examiner(s): Dr. T. Dowling Internal Examiner(s): Dr. J. Creagh
Instructions: Answer THREE questions. Note 120 marks maximum.
Duration: 2 Hours
Sitting: Semester 2 2011
Requirements for this examination:
Note to Candidates: Please check the Programme Title and the Module Title to ensure that you have received the correct examination. If in doubt please contact an Invigilator.
Q1. (a) i) Explain LCG generators. Use m = 9, a = 7, c = 5 and Z 0 = 8 to generate a sample sequence. (8 marks) ii) Give an algorithm for generating a random number U(0,1) using the LCG generator method. (7 marks) (b) PMMLCG generators are used by simulation packages such as SIMAN. Evaluate such a decision. Clear justifications required. (7 marks) (c) Consider the SIMAN call of DISC(0.2, 8, 0.3, 9, 0.5, 10, 0.95, 11, 1.0, 12); Explain this example. Write an algorithm (or C/C++ program) for implementing this call. (8 marks) (d) Consider the SIMAN call of EXPO(4.0); Clearly explain the functionality of this call in your own words. Write an algorithm (or C/C++ program) for implementing this call. (10 marks) [40 marks]
Q2. (a) A modeller must firstly gather data before deciding how to model input to simulation. Evaluate the 4 principal approaches to generating Input to Simulation from the viewpoint of a modeller. (10 marks) (b) Mean square error and p-value are useful measures when deciding on which theoretical distribution to use. Discuss. (8 marks) (c) For theoretical distributions, the analysis of the observed data should take many factors into account. Explain, discuss and give recommendations for a modeller when outliners and multiple-peaks are discovered. (12 marks) (d) From a modelling viewpoint, explain the usefulness of each of the following distributions: Weibull Exponential Beta Erlang (10 marks) [40 marks]
Q4. (a) Explain the 3 key areas of simulation model validation. Justify why you think each area is so important to simulation model validation. (8 marks) (b) For analysis of non-terminating systems, compare replication and batching observations methods used in estimating the variance of the mean. (7 marks) (c) The simulation process specified by Pegden et. al. requires several stages/steps. Outline each of the stages for the model design phase. In your opinion, why is each of these stages needed? Justify your answer. (9 marks) (d) Consider a simple system to model the performance of a sales machine. Customers arrive at a triangular rate of between 2 and 7 minutes with an average of 4 minutes. The sales machine servers at an exponential rate of 3 minutes. Server down time must also be taken into account. Choose your own times regarding the server down times. Develop a SPN model. (The diagram is only required) Develop an event graph model. (The diagram is only required) Compare these two solutions. Indicate your preferred solution and explain the reasons for your choice. (16 marks) [40 marks]
Q5. (a) i) When should a simulation developer consider using parallel and distributed simulation? (3 marks) ii) From a parallel and distributed simulation viewpoint, evaluate the characteristics of analytic simulation and virtual environments. (6 marks) (b) i) Explain the concept of an “anti-message” for optimistic PDES. (4 marks) ii) In your opinion, what are the benefits of fossil collection with PDES? (4 marks) (c) When writing a HLA simulation program, when should HLA objects be used and when should interactions be used? Examples and comparisons are required. (8 marks) (d) HLA enables optimistic execution among a collection of optimistic federates. i) Give the algorithm/code structure for the main loop for a typical Time Warp federate. (7 marks) ii) What are the key callbacks that are involved for a Time Warp federate program? Outline/explain how these callbacks should be designed, giving an example. (8 marks) [40 marks]
