2^k Factorial Design

Definition

A $2^k$ Factorial Design is an experimental strategy where $k$ factors are tested at exactly two levels each (usually “Low” and “High”). It is the most efficient method for screening a large number of variables to identify those with the most significant impact on system performance.

• How to read: “Two to the k factorial design; k factors.”
• Meaning: Every factor is binary (low/high); all $2^k$ corner combinations of the design space are tested to map main and interaction effects.

Why It Matters

This is the ‘efficiency king’ of experimental design. It allows researchers to test the effects of multiple factors (and their interactions) with a minimal number of runs, providing the most information for the least effort in complex system optimization.

Core Concepts

• Factors ($k$): The variables being tested.
  • How to read: “k factors.”
  • Meaning: $k$ is the count of independent variables you manipulate in the experiment.
• Levels (-1, +1): The low and high settings for each factor.
  • How to read: “Levels negative one and positive one.”
  • Meaning: Coded low/high values; center the design at zero and scale each factor to $\pm 1$ for symmetric analysis.
• Total Runs: $2^k$ (e.g., 3 factors require $2^3 = 8$ distinct scenarios).
  • How to read: “Two to the k; two cubed equals eight.”
  • Meaning: Run count grows exponentially with $k$; use for screening when $k$ is modest (typically $\leq 5$).
• Main Effect: The average change in the response variable when a specific factor is moved from Low to High.
• Interaction Effect: When the impact of one factor depends on the level of another factor (e.g., adding a machine only helps if you also add an operator).
• Efficiency: Allows for the study of interactions, which one-factor-at-a-time (OFAT) testing cannot detect.

Connected Concepts

• Experimental Design in Simulation
• Interaction Effects
• Optimization