Design of Simulation Experiments Overview

Simulation models often have many input factors and determining which ones have a significant impact on performance measures (responses) of interest can be a truly daunting task. The common approach of changing one factor at a time is very often incorrect, because for many models factors interact to impact on the responses. In this simulation course, we give a comprehensive presentation of design of experiments (DOE) specifically for simulation modeling, whose major goal is to determine which factors have the greatest effect on the responses, and to do so with the least amount of simulating. Other important uses of DOE are to develop a metamodel (a simplified model of the simulation model) based on the important factors to predict the model response for factor-level combinations that were not actually simulated due to execution- or setup-time constraints or to find the factor combination that optimizes the simulation response.

We discuss a simple and widely applicable approach for determining significant factors in the context of simulation modeling, whereas commonly used methods based on classical statistics (i.e., ANOVA) make assumptions that are rarely satisfied in practice. Students will analyze experimental data in class using a leading DOE software package, and be prepared to apply their knowledge the following week at work.

Each attendee will receive a copy of the book Simulation Modeling and Analysis (5th Edition, McGraw-Hill, January 2014) by Dr. Averill M. Law as part of their registration fee – this book is widely considered to be the “bible” of simulation with more than 152,000 copies in print and 14,000 citations. Versions of this seminar have been presented to Lockheed Martin, Middle East Technical University/Roketsan (Turkey), NSA, Sasol Technology (South Africa), and the U.S. Navy.

Outline

What You Will Learn:

  1. Factorial Designs for Determining Important Model Factors
  2. Fractional Factorial Designs for Reducing the Required Number of Simulation Runs
  3. Classical Metamodels Based on Central-Composite Designs and Regression for Prediction or Optimization
  4. Simulation Metamodels Based on Space-Filling Designs and Kriging (Gaussian Process Modeling)
  5. Optimal or Computed-Generated Designs
  6. Critical Dangers of Using Standard Experimental Designs and Analyses for Simulation Modeling
  7. Commercial Software for DOE That Are Appropriate for Simulation Modeling
  8. Numerous Examples and In-Class Exercises to Illustrate the Mechanics and Applications of DOE