MSCI302-09S2 (C) Semester Two 2009

Probabilistic Operations Research Models

28 points

Details:
Start Date: Monday, 13 July 2009
End Date: Sunday, 15 November 2009
Withdrawal Dates
Last Day to withdraw from this course:
  • Without financial penalty (full fee refund): Sunday, 26 July 2009
  • Without academic penalty (including no fee refund): Sunday, 11 October 2009

Description

Applied stochastic processes. Queueing models and their use. Discrete event simulation using commercial packages. A brief introduction to continuous-systems and Monte Carlo simulation. Business forecasting. A mainstream course for MSCI/OR majors.

Learning Outcomes

To cover these topics from an MS/OR practitioner's viewpoint.  To ensure that you know what the basic theory and applications of Markov chain models are, and are capable of applying and fitting a queueing model to real data, can  set up and run a simple simulation experiment, and can apply extrapolative forecasting techniques.

Prerequisites

(1) MSCI204; (2) MSCI210 or 22 points of 200-level courses in STAT; (3) any one of COSC121, ACIS125, AFIS123, ENEL206, MATH171, MATH282, or
any course involving an appropriate level of computer programming, as approved by the Head of Department.

Restrictions

MSCI310, MSCI311, MSCI312

Course Coordinator / Lecturer

Don McNickle

Assessment

Assessment Due Date Percentage  Description
Assignment 1 21 Aug 2009 16% Assignment 1
Assignment 2 18 Sep 2009 22% Assignment 2
Assignment 3 19 Oct 2009 12% Assignment 3
Final Examination 50% Final Examination


One A4 page of notes (2 sides) is allowed for the exam.

Textbooks / Resources

Recommended Reading

Operations Research: Applications and Algorithms ;

Banks, Jerry. et al; Getting started with GPSS/H ; 2nd ed; Wolverine Software, 1995.

Makridakis, Spyros G. , Wheelwright, Steven C., Hyndman, Rob J; Forecasting : methods and applications ; 3rd ed. ; Wiley, 1998.

Notes

Grading:
The marks for the assessment items will be standardised before a final grade is determined.  You should not regard 50% as a pass mark.

COURSE NOTES
Perhaps one-third of Management Science techniques are probabilistic or stochastic. In addition many other problems require estimation and fitting of probabilistic models to actual data.  

What I will do:
I will ensure that you have at least three weeks to complete the assignments. It is normal practice to return assignments by placing them out in boxes or bringing them along to class, so that you can get them as soon as possible. If you object to your work being returned in this way you should indicate this in writing and your work will be returned separately by the departmental secretary.

What you must do:
Unless there are unexpected circumstances beyond your control the assignment must be handed in by the due date if you want it to count towards your grade. Since it will count towards your final grade it must also be entirely your own work, except for the pairs provision in assignment 1 i.e. no plagiarism, collusion or copying. You should do about 11 hours of study per week. (This course is half a normal semester load, assuming you work 30 hours per week, and we have 4 contact hours that leaves 11 hours).

TOPICS
This course aims to provide students majoring in Management Science with an appropriate background in the methods and theory of stochastic processes, queueing theory, computer simulation and forecasting. The course will cover the following material.

1.  Stochastic Processes
Elementary stochastic processes, probability distributions, expectations, Poisson processes, Markov chains

2.  Queueing Theory
Markovian queueing models. How to derive steady-state characteristics of these models.  M/G/1 queues, phase-type approximations. The assignment is to fit a theoretical model to an actual queue. This has benefits beyond this section since the analysis of time-dependent data is common to simulation and other O.R. techniques as well.

3. Computer Simulation
Computer simulation is one of the most widely used Management Science techniques. The object is to present simulation from a skilled practitioner's viewpoint. The principal simulation approaches to discrete-event simulation programming, generation of random variates, strategies for running simulations to obtain valid results, statistical analysis of simulation output. A survey of simulation packages. Short introductions to continuous systems simulation and Monte-Carlo methods. Two simulation packages, I will be asssuing you have studied Simul8 in some detail in MSCI204, but GPSS/H will be covered  extensively here so you see both the GUI and language approaches to simulation programming.

4. Forecasting
Extrapolative versus Explanatory models. Extrapolative forecasting techniques including classical decomposition, exponential smoothing, ARIMA models, neural networks, rule-based and business forecasting. Judgemental forecasting and biases. We use the Forecast Pro package extensively.

Departmental Academic Policies
If you want a hard copy of this document, please ask the course co-ordinator. The Department assumes that you have read this document. You should also read the “Information related to courses and assessment” on page 372 of the Enrolment Handbook 2009 (also in UC Calendar under “General Course and Examination Regulations”).

Coversheets - Group and Individual

Indicative Fees

Domestic fee $1,054.00

International fee $4,433.00

* All fees are inclusive of NZ GST or any equivalent overseas tax, and do not include any programme level discount or additional course-related expenses.

For further information see Management, Marketing and Tourism .

All MSCI302 Occurrences

  • MSCI302-09S2 (C) Semester Two 2009