CS 555: Analytic Models and Simulation of Computer Systems

Objectives

• The course presents analytic and simulation techniques used for performance evaluation and modeling of computer systems and networks.
• Fundamentals of probability theory are discussed first followed by rigorous development of different probability distributions and queueing systems.
• Various aspects of discrete-event and Monte Carlo simulations are studied including the methodology of statistical evaluation of simulation results.
• The covered material is illustrated with many practical examples.

Prerequisites

• CS 450.

Syllabus

• Methodology of modeling and performance evaluation
• Review of probability theory
• Probability measures
• Conditional probability
• Combinatorial analysis
• Random variables
• Probability density function
• Probability mass function
• Cumulative distribution function
• Exponential distribution
• Poisson distribution
  • Queueing models
• Birth-and-death equations
• Single Markovian queueing systems
• The Erlang B formula
• The Equivalent Random Method
• The Gauss-Seidel iteration
• Overrelaxation method
  • Simulation
    • Discrete event simulation
    • Monte Carlo simulation
    • Generation of probability distributions
    • Statistical evaluation of simulation results
    • Goodness-of-fit tests
  • Applications
• Packet switching
• Circuit switching
• Teletraffic engineering

Edited March 2006 (html, css checks)