01 / 21

Function Model
Construction & Application

Learn to build mathematical models from data and context, selecting the right function type to represent real-world relationships.

AP Precalculus

What Is a Function Model?

A function model is a mathematical function chosen to describe a real-world relationship between quantities. We use data, context clues, and known function behaviors to select and build these models.

Input
The independent variable: time, distance, price, etc.
Output
The dependent variable: population, revenue, height, etc.
Model
The function \( f(x) \) that maps every input to an output.

The Model-Building Cycle

Observe Data Choose Type Build Model Validate / Apply Revise

Why Do Models Matter?

1
Prediction: Estimate future values beyond the observed data (extrapolation).
2
Interpolation: Estimate values between known data points.
3
Optimization: Find maximum or minimum values of a quantity.
4
Communication: A compact formula conveys a relationship more precisely than a table of numbers.

Core Function Families

Linear
\( f(x) = mx + b \)
Constant rate of change
Quadratic
\( f(x) = ax^2 + bx + c \)
Changing rate of change (linear)
Exponential
\( f(x) = a \cdot b^x \)
Constant proportional change
Logarithmic
\( f(x) = a + b\ln(x) \)
Rapid then slowing growth

Signature Behaviors

x y Linear Quadratic Exponential Log

Linear vs. Exponential

Linear Exponential
General Form \( f(x) = mx + b \) \( f(x) = a \cdot b^x \)
Rate of Change Constant Proportional to value
Data Test Equal \(\Delta y\) over equal \(\Delta x\) Equal ratio \(\frac{y_2}{y_1}\) over equal \(\Delta x\)
Long-term Behavior Grows without bound, steadily Grows without bound, accelerating

Identifying the Right Model

1
Compute first differences \( \Delta y = y_{n+1} - y_n \). If roughly constant: linear.
2
Compute ratios \( \frac{y_{n+1}}{y_n} \). If roughly constant: exponential.
3
Compute second differences \( \Delta^2 y \). If roughly constant: quadratic.
4
Context clues: Look for keywords like "doubles every," "added per year," or "half-life."

Step-by-Step Model Construction

1
Read the context. Identify the independent and dependent variables.
2
Analyze the data. Compute differences, ratios, or second differences.
3
Select the function family whose signature matches the data pattern.
4
Determine parameters. Use known points to solve for \( m, b \) or \( a, b \) etc.
5
Validate. Check the model against additional data points or expected behavior.

Follow the salt through a working mixing rig

A stirred transparent tank has one upper inlet and one lower outlet with a valve. The exact diagram links its 100 L salt balance to a computed approach to equilibrium.With the outlet closed, salt mass is 6t g and water volume is 50 + 2t L. Their ratio C(t) = 3t/(25 + t) g/L increases toward the incoming concentration of 3 g/L.

Filling-only concentration model

With the outlet closed, salt mass is 6t g and water volume is 50 + 2t L. Their ratio C(t) = 3t/(25 + t) g/L increases toward the incoming concentration of 3 g/L.With the outlet closed, salt mass is 6t g and water volume is 50 + 2t L. Their ratio C(t) = 3t/(25 + t) g/L increases toward the incoming concentration of 3 g/L.

Example 1: Linear Model

A water tank drains at a steady rate. At \( t = 0 \) hours it holds 200 gallons, and at \( t = 5 \) hours it holds 75 gallons. Build a model for \(W(t)\).

1
Rate of change: \( m = \frac{75 - 200}{5 - 0} = \frac{-125}{5} = -25 \) gal/hr.
2
Initial value: \( W(0) = 200 \), so \( b = 200 \).
3
Model: \( W(t) = -25t + 200 \).

Example 1: Graph

t (hours) W (gallons) 0 2 4 6 8 0 50 100 150 200 (0, 200) (5, 75) (8, 0)

Example 2: Exponential Model

A bacteria culture starts with 500 cells and doubles every 3 hours. Build a model for the population \(P(t)\).

1
Initial value: \( P(0) = 500 \), so \( a = 500 \).
2
Growth factor: Doubles every 3 hours, so \( b = 2 \) per 3-hour period: \( P(t) = 500 \cdot 2^{t/3} \).
3
Validate: \( P(3) = 500 \cdot 2^1 = 1000 \) ✓, \( P(6) = 500 \cdot 4 = 2000 \) ✓.

Example 2: Graph

t (hours) P (cells) 0 3 6 9 12 0 2k 4k 6k 8k (0, 500) (3, 1k) (6, 2k) (9, 4k)

Example 3: Quadratic from Data

A ball is launched upward. Its height \(h\) (meters) at time \(t\) (seconds) is recorded:

\( t \) 01234
\( h(t) \) 227424742
\( \Delta h \) 25155-5
\( \Delta^2 h \) -10-10-10

Common Mistakes

✕ Using ratios on zero outputs
If any \(y\)-value is 0 or negative, the ratio test for exponential is undefined. Check differences first.
✕ Ignoring domain restrictions
A model like \( P(t) = 500 \cdot 2^{t/3} \) may only be valid for \( t \geq 0 \). Don't extrapolate blindly.
✕ Confusing growth factor with rate
If a population grows 5% per year, \( b = 1.05 \), not \( 0.05 \). The base includes the original amount.

Summary of Function Models

Model Form Key Signal Context Clue
Linear \( mx + b \) Constant \( \Delta y \) "per year," "each month"
Quadratic \( ax^2 + bx + c \) Constant \( \Delta^2 y \) "projectile," "area"
Exponential \( a \cdot b^x \) Constant ratio "doubles," "half-life"
Logarithmic \( a + b\ln(x) \) Fast then slow growth "diminishing returns"

Question 1 of 5

A dataset has outputs 10, 20, 40, 80, 160 for equally spaced inputs. Which function type best models this data?

Question 2 of 5

A car depreciates from $20,000 by 15% each year. Which model represents its value \(V(t)\) after \(t\) years?

Question 3 of 5

The second differences of a data set are approximately constant at 6. What is the leading coefficient \(a\) in the quadratic model \(f(x) = ax^2 + bx + c\)?

Question 4 of 5

An exponential function passes through \( (0, 4) \) and \( (2, 36) \). What is the value of \(b\) in \(f(x) = a \cdot b^x\)?

Question 5 of 5

A linear model gives \(f(10) = 50\) and \(f(15) = 75\). What is \(f(20)\)?