Comparing and modeling functions Knowledge-Based Study

Study notes

1. Fundamental Patterns of Function Growth

   When analyzing tables of values or function rules, the relationship between successive outputs for evenly spaced inputs reveals the underlying family of the function:

   • Linear Functions: Exhibit a constant first difference. The function changes by adding or subtracting a fixed quantity: $f(x) = mx + b$.
   • Quadratic Functions: Exhibit a constant second difference. The first differences change at a constant rate: $f(x) = ax^2 + bx + c$.
   • Exponential Functions: Exhibit a constant ratio (multiplicative factor) between consecutive outputs: $f(x) = a \cdot b^x$, where $b > 0$ and $b \neq 1$.

   Recognizing these distinct patterns allows us to choose the most mathematically faithful model for a given set of data points.

   The Average Rate of Change

   The average rate of change measures how much the function output changes per unit change in input over an interval $[a, b]$:

   Average Rate of Change: $\frac{f(b) - f(a)}{b - a}$

   Different families of functions demonstrate contrasting behaviors across successive intervals of equal length:

   • For a linear function, the average rate of change is constant across every possible interval and equals the slope $m$.
   • For a quadratic function with $a > 0$, the average rate of change increases at a constant linear rate as the interval shifts to the right.
   • For an increasing exponential function ($b > 1$), the average rate of change increases exponentially as the interval shifts to the right, scaling by the factor $b$ over shifted intervals of unit length.

   End Behavior and Dominance

   A central property in comparative function analysis is the long-term behavior of different function families as the independent variable $x$ grows without bound ($x \to \infty$):

   For any base $b > 1$, any positive constant $a$, and any exponent $n > 0$, the exponential function $f(x) = a \cdot b^x$ will eventually exceed any polynomial function $g(x) = c \cdot x^n$ for sufficiently large values of $x$, regardless of how large the coefficient $c$ or degree $n$ may be.

   • Linear growth (degree 1) is eventually surpassed by quadratic growth (degree 2).
   • Polynomial growth of any fixed degree is eventually surpassed by exponential growth.
   • Although a linear or polynomial function may produce greater outputs on a short initial interval, the multiplicative nature of exponential growth guarantees eventual dominance.

   Selecting and Validating Models

   To determine the most appropriate model when given numerical or verbal representations, apply the following systematic process:

   1. Analyze consecutive differences: Compute $\Delta y = y_{k+1} - y_k$ for equal steps $\Delta x$. If $\Delta y$ is constant, select a linear model.
   2. Analyze consecutive ratios: Compute $\frac{y_{k+1}}{y_k}$. If this quotient remains constant, select an exponential model.
   3. Analyze second differences: If $\Delta(\Delta y)$ is constant, select a quadratic model.
   4. Contextual verification: Ensure that the model matches theoretical boundary conditions, such as non-negative constraints, vertical intercepts, or horizontal asymptotes.

   By systematically comparing rates of change and output behavior, one can select, construct, and defend the optimal functional model.