Definition
Experimental Design is the systematic process of determining which factors (inputs) to vary and how to measure the resulting response (outputs) to answer specific research questions with the minimum number of simulation runs.
Why It Matters
Experimental design is the systematic process of maximizing the “signal” of your research while minimizing the computational “noise.” It leverages the principle of parsimony to identify the most critical factors of a complex system with the minimum number of runs, ensuring that your optimization efforts are both efficient and accurate.
Core Concepts
- Factors (Independent Variables): The inputs being changed (e.g., number of machines, arrival rate).
- Responses (Dependent Variables): The outputs being measured (e.g., throughput, cost).
- Levels: The specific values assigned to each factor (e.g., 2, 4, or 6 machines).
- Full Factorial Design: Testing every possible combination of every factor at every level.
- Fractional Factorial Design: Testing a subset of combinations to save time, focusing on main effects and key interactions.
- Goal: To identify which factors have the greatest impact on the system and to find the optimal “Operating Point.”