Mathematical Modeling and Applications
The Mathematical Modeling and Applications domain tests the full modeling cycle: translating a real situation into mathematics, computing, interpreting results in context, and judging whether the model is reasonable.
The cycle and assumptions
Formulate by defining variables and stating assumptions; compute with an appropriate model; interpret each parameter and output in the original units; validate against data or reasonableness. Every model idealizes, so state what is being ignored.
Common model families
Linear y = mx + b fits constant absolute change. Exponential y = abˣ fits constant percent change, with b = 1 + r and continuous form y = ae^{kt}. Quadratic models suit optimization, with the vertex at x = −b/(2a). Logistic models capture bounded growth. Choose by inspecting differences (constant differences indicate linear) versus ratios (constant ratios indicate exponential).
Worked example
A town of 12,000 grows 3% per year. Model P(t) = 12,000(1.03)ᵗ. Doubling time solves 2 = 1.03ᵗ, so t = ln 2 / ln 1.03 ≈ 23.4 years. Interpreting the base: each year's population is 1.03 times the previous year's, not 3 more people.
Traps
Extrapolating far beyond the data; confusing percentage change with percentage points; reporting a model output without units; assuming correlation establishes causation; trusting a high R² when residuals show systematic curvature.