CS 437 Computer Simulation of Complex Systems
Objectives
Many processes, useful in Computer Science, Finance, Operations Research, Business, and science are sufficiently complex to require analysis by computer simulation methods. These include queues, networks, inventory models, and stochastic models for stock prices and interest rates. Such models can be used for predictive purposes, risk management or to price financial derivatives.
This course is designed to teach the process of building such simulation models, improving their efficiency, implementing and analyzing the results.
Intended Audience
This course is for students who are not CS majors.
Related Courses
Prerequisites: (CS 126/124/114 or 134 or 136 or SYDE 221/322) and ( STAT 231/241 or SYDE 214); Not open to Computer Science or General Mathematics students.
Antirequisites: CM361/STAT 341, CS 457
Cross-listed as: STAT 340.
References
TBA.
Schedule
3 hours of lectures per week. Normally available in Winter and Spring.
Outline
Introduction to Simulation (6 hours)
Why simulate? Examples of discrete event simulations. Simulating simple games and queuing systems. Performance measures. Introduction to simulation languages (e.g. Simul8, Matlab, Excel).
Model Building and Validation (3 hours)
How to build a model that accurately represents a real-world phenomena. A simple model for stock prices.
Generation of Uniform and Non-Uniform Random Numbers (8 hours)
Algorithms for generating random numbers under a variety of distributions. What to watch for in a poor random number generator. Testing a uniform random number generator.
Variance Reduction Techniques and Applications (8 hours)
Various methods for improving the efficiency of a simulation are discussed including the use of control and antithetic random numbers, stratified sampling, importance sampling, combining various methods. Applications to pricing a call option. How to get more bang for your CPU time. Elements of Quasi Monte Carlo.
Sensitivity of Simulations (2 hours)
How to estimate derivatives using a simulation. What if the underlying parameter values change by a little or a lot?
Analysis of Output (2 hours)
How to analyse and display the results from simulations.
Other Uses of Simulation (2 hours)
The bootstrap. Tests conducted by simulation.