Function Model Selection and Assumption Articulation
Why this matters
Imagine your school's robotics club, the "Circuit Breakers," wants to sell custom t-shirts to fund a trip to a competition in Dallas. You need to figure out pricing. A local print shop charges a $100 setup fee plus $8 per shirt. You also know that if you price the shirts too high, fewer people will buy them. How do you model your costs? How do you model your potential profit? Is one type of equation enough?
Choosing the right mathematical tool—the right type of function—is the first step to answering these questions and making a smart decision. In this lesson, we'll learn how to analyze a situation, select the best function model, and clearly state the assumptions and limitations that come with it. This is how math gets used to make real-world decisions.
Concept overview
flowchart TD
A[Analyze scenario or data] --> B{Is rate of change constant?};
B -- Yes --> C[Use Linear Model];
B -- No --> D{Symmetric data? <br/> One max/min?};
D -- Yes --> E[Use Quadratic Model];
D -- No --> F{Multiple turning points?};
F -- Yes --> G[Use Higher-Degree Polynomial];
F -- No --> H{Do rules change at specific points?};
H -- Yes --> I[Use Piecewise Model];
H -- No --> J[Consider other function types];
Core explanation
Hello! I'm Saavi, and I'm so glad you're here. Today, we're moving from pure calculation into the art of mathematical modeling. Think of it like this: you have a toolbox filled with different functions (lines, parabolas, and more). Your job is to look at a real-world problem and pick the perfect tool for the job.
Part 1: Choosing Your Function Model
How do you know which function to pick? You look for clues in the problem description or the data.
Clue #1: A Constant Rate of Change → Linear Function
If a quantity changes by the same amount for each unit increase in another, you're looking at a linear function. The classic form is f(x) = mx + b.
- The Telltale SignThe rate of change is constant.
- ExampleA car service in Chicago charges a $5 flat fee and then $2.50 per mile. The cost increases by exactly $2.50 for every single mile driven. That constant rate is your slope,
m. The cost function would beC(d) = 2.50d + 5.
Clue #2: A Symmetrical Shape or Changing Rate → Quadratic Function
Quadratic functions are perfect for two common scenarios. They make the U-shape of a parabola, f(x) = ax^2 + bx + c.
- The Telltale Signs:
- Data that is roughly symmetrical, rising to a single maximum point and then falling (or falling to a minimum and rising). Think of a basketball shot: it goes up, reaches a peak height, and comes back down.
2. A rate of change that is *itself* changing at a constant rate. This is where most students get stuck. For a linear function, the *value* changes. For a quadratic, the *rate of change* changes—but it does so linearly.- Geometric Context: Area often leads to a quadratic model. If you have a square photo with side length
s, its area isA = s^2. If you're building a rectangular dog run, the area will involve multiplying two dimensions, often leading to a quadratic.
Clue #3: Multiple Bumps or 3D Space → Cubic & Higher-Degree Polynomials
When things get more complex, you may need a polynomial with more curves.
-
The Telltale Signs:
- 1Multiple "turning points"The graph has several local maximums or minimums. Think of a roller coaster with multiple hills and valleys.
- 2VolumeGeometric problems involving three dimensions, like the volume of a box, often result in cubic functions. If you make a box by cutting squares of side length
xfrom the corners of a piece of cardboard, the volume will be a cubic function ofx.
-
The Finite Differences Test: This is a powerful tool when you have a table of data.
- Calculate the differences between consecutive y-values (the "first differences"). If they are constant, the function is linear.
- If not, calculate the differences between those differences (the "second differences"). If they are constant, the function is quadratic.
- If the nth differences are constant, the function is a polynomial of degree n.
-
The
n+1Points Rule: A set ofn+1points with unique x-values can be modeled perfectly by a polynomial of degree at mostn. So, if you have 4 data points, you can find a cubic function (degree 3) that passes through all of them.
Clue #4: Different Rules for Different Intervals → Piecewise-Defined Function
Sometimes, a situation doesn't follow one single rule. The rules change depending on the input value.
- The Telltale SignYou see phrases like "up to," "after that," "for the first 50," or "for any amount over 100."
- ExampleYour cell phone plan. You might pay $40 for the first 10 GB of data, but then $10 for each additional GB you use. The cost function has two different pieces: one for when you use 10 GB or less, and another for when you use more.
/ 40, if 0 <= x <= 10
C(x) = {
\ 40 + 10(x - 10), if x > 10Part 2: Stating Your Assumptions and Restrictions
Picking the function is only half the battle. A model is a simplification of reality. You must state the rules you're playing by. This is what separates a guess from a mathematical model.
Assumptions: The "What Ifs"
Assumptions are conditions you accept as true for your model to work, even if they aren't perfect in the real world.
- What is consistent?In our t-shirt example, we assume the cost of a blank shirt from the supplier ($8) doesn't suddenly jump to $9 halfway through our order.
- How do quantities change together?When modeling profit vs. price, we might assume a smooth, predictable relationship where raising the price always lowers demand. In reality, a price jump from $19.99 to $20.00 might have a bigger psychological impact than a jump from $21.00 to $21.02. Our model simplifies this.
Restrictions: The "Rules of the Game"
Restrictions define the valid inputs (domain) and outputs (range) for your model based on its context.
- Domain Restrictions (Inputs)What
xvalues make sense?- Contextual Clue: If
xrepresents the number of t-shirts sold,xcannot be negative. It also can't be a fraction; you can't sell 23.7 shirts. So, the domain is non-negative integers.x >= 0,xis an integer. - Mathematical Clue: If your model is
f(x) = 1 / (x-2), the domain must excludex=2to avoid division by zero.
- Contextual Clue: If
- Range Restrictions (Outputs)What
yvalues make sense?- Contextual Clue: If
h(t)models the height of a thrown baseball, the height can't be negative (unless you're on a cliff). - Rounding: If your function calculates the number of buses needed for a field trip and gives you 4.3, the practical range value is 5. You must round up. The range is restricted to positive integers.
- Contextual Clue: If
Articulating these points shows you understand that math isn't just about numbers; it's about using numbers to describe the world thoughtfully.
Worked examples
Let's put these ideas into practice with a couple of examples.
The Softball Toss
Priya is on her school's softball team. She tosses a ball straight up into the air. Her friend Marcus records the height of the ball at different times.
| Time (s) | Height (ft) |
|---|---|
| 0 | 5 |
| 0.5 | 17 |
| 1.0 | 21 |
| 1.5 | 17 |
| 2.0 | 5 |
Problem: A. What type of function would best model the height of the ball as a function of time? B. What are the key assumptions and restrictions for this model?
Solution:
Part A: Identify the Function Type
- 1Analyze the DataLook at the
Heightcolumn. The ball starts at 5 ft, goes up to a peak of 21 ft, and then comes back down. Notice the symmetry: the height at 0.5s (17 ft) is the same as the height at 1.5s. The height at 0s (5 ft) is the same as at 2.0s. - 2Connect to Function PropertiesThis symmetry around a single maximum value (21 ft at t=1.0s) is the classic signature of a quadratic function.
- 3Check Finite Differences (Optional but good practice)
- First differences:
17-5=12,21-17=4,17-21=-4,5-17=-12. These are not constant. - Second differences:
4-12=-8,-4-4=-8,-12-(-4)=-8. The second differences are constant (-8). This confirms it's a quadratic model.
- First differences:
Answer for A: A quadratic function is the most appropriate model because the data is symmetric around a single maximum value, and the second differences are constant.
Part B: State Assumptions and Restrictions
- 1AssumptionsWhat are we ignoring?
- We assume there is no wind resistance or other atmospheric effects. We're modeling a "perfect" toss in a vacuum.
- We assume gravity is the only force acting on the ball after it leaves Priya's hand.
- 2RestrictionsWhat are the sensible limits on our variables?
- Domain (Time
t): Time can't be negative, sot >= 0. The ball also hits the ground (or is caught) at some point. Based on the data, the flight seems to last a little over 2 seconds. A reasonable domain might be0 <= t <= 2.0if we assume it's caught at the same height it was thrown. - Range (Height
h): The height cannot be negative. The minimum height in the data is 5 ft (the release point), and the maximum is 21 ft. So, a reasonable range is5 <= h <= 21.
- Domain (Time
The Bulk Discount
A small business in Atlanta sells custom-printed coffee mugs. For orders of 50 mugs or fewer, the price is $10 per mug. If a customer orders more than 50 mugs, the price for the entire order drops to $8 per mug to encourage large orders.
Problem:
A. Construct a function C(x) that models the total cost for an order of x mugs.
B. State any restrictions on the domain of your function.
Solution:
Part A: Construct the Function
- 1Identify the "Break Point"The rule for the cost changes at
x = 50. This is a huge clue that we need a piecewise-defined function. - 2Define the First PieceFor orders of 50 or fewer (
x <= 50), the cost is simply $10 times the number of mugs.Cost = 10xfor0 < x <= 50.
- 3Define the Second PieceFor orders of more than 50 (
x > 50), the cost for the entire order is $8 times the number of mugs.Cost = 8xforx > 50.
- 4Combine into a Piecewise Function
/ 10x, if 0 < x <= 50 C(x) = { \ 8x, if x > 50
Part B: State Domain Restrictions
- 1Consider the Context
xrepresents the number of coffee mugs. - 2Apply Logic
- Can you order a negative number of mugs? No. So,
x > 0. - Can you order a fraction of a mug? No. So,
xmust be an integer.
- Can you order a negative number of mugs? No. So,
- 3State the Restriction ClearlyThe domain of
C(x)is the set of positive integers.
This example shows how a real-world business rule directly translates into the structure of a piecewise function. Forgetting the integer restriction is a common oversight.
Try it yourself
Ready to try on your own? Here are a couple of scenarios to test your modeling skills.
1. The Garden Box
Carlos is building a rectangular garden box with an open top. He has a piece of wood that is 12 feet long and 8 feet wide. He plans to cut out identical squares of side length x from each of the four corners and then fold up the sides.
- Your task:
- Write a function
V(x)that represents the volume of the box. What type of function is it? - What are the practical restrictions on the domain of
x? (Hint: Think about the dimensions of the wood. How big can the cutoutxbe before the sides disappear?)
- Write a function
2. The Phone Bill
Maya's phone plan in Boston has the following structure:
-
A flat fee of $35 per month, which includes up to 5 GB of data.
-
For any data usage over 5 GB, she is charged an additional $8 per GB.
-
Your task:
- Construct a piecewise-defined function
C(g)that models Maya's total monthly cost for usingggigabytes of data. - What is the cost for using 4.5 GB? What about 7 GB?
- Construct a piecewise-defined function
Give these a shot! The key is to translate the words into mathematical relationships and then think about the real-world limits.
Practice — 8 questions
In simple terms, this topic is about choosing the right type of function to represent a real-world situation and stating the rules and assumptions that make your model work.
/ 40, if 0 <= x <= 10
C(x) = {
\ 40 + 10(x - 10), if x > 10- 1.13.A: Identify an appropriate function type to construct a function model for a given scenario.
- 1.13.B: Describe assumptions and restrictions related to building a function model.
- 1.13.A.1
- Linear functions model data sets or aspects of contextual scenarios that demonstrate roughly constant rates of change.
- 1.13.A.2
- Quadratic functions model data sets or aspects of contextual scenarios that demonstrate roughly linear rates of change, or data sets that are roughly symmetric with a unique maximum or minimum value.
- 1.13.A.3
- Geometric contexts involving area or two dimensions can often be modeled by quadratic functions. Geometric contexts involving volume or three dimensions can often be modeled by cubic functions.
- 1.13.A.4
- Polynomial functions model data sets or contextual scenarios with multiple real zeros or multiple maxima or minima.
- 1.13.A.5
- A polynomial function of degree n models data sets or contextual scenarios that demonstrate roughly constant nonzero nth differences.
- 1.13.A.6
- A polynomial function of degree n or less can be used to model a graph of n + 1 points with distinct input values.
- 1.13.A.7
- A piecewise-defined function consists of a set of functions defined over nonoverlapping domain intervals and is useful for modeling a data set or contextual scenario that demonstrates different characteristics over different intervals.
- 1.13.B.1
- A model may have underlying assumptions about what is consistent in the model.
- 1.13.B.2
- A model may have underlying assumptions about how quantities change together.
- 1.13.B.3
- A model may require domain restrictions based on mathematical clues, contextual clues, or extreme values in the data set.
- 1.13.B.4
- A model may require range restrictions, such as rounding values, based on mathematical clues, contextual clues, or extreme values in the data set.
flowchart TD
A[Analyze scenario or data] --> B{Is rate of change constant?};
B -- Yes --> C[Use Linear Model];
B -- No --> D{Symmetric data? <br/> One max/min?};
D -- Yes --> E[Use Quadratic Model];
D -- No --> F{Multiple turning points?};
F -- Yes --> G[Use Higher-Degree Polynomial];
F -- No --> H{Do rules change at specific points?};
H -- Yes --> I[Use Piecewise Model];
H -- No --> J[Consider other function types];
Read what Saavi narrates
Hi there, I'm Saavi. Let's talk about something that's at the very heart of applying math to the real world: choosing the right tool for the job.
Imagine your school's robotics club wants to sell custom t-shirts to fund a trip. You find a print shop that charges a one-hundred-dollar setup fee plus eight dollars per shirt. You also know that if you price the shirts too high, fewer people will buy them. How do you model your costs? How do you figure out the best price for maximum profit?
This is what today's lesson is all about. We're learning to be detectives. We'll look at clues in a problem to pick the best function—is it a straight line, a parabola, or something else? Then, we'll act like good scientists and state the "ground rules"—the assumptions and restrictions—that make our model make sense.
Let's walk through an example together. Imagine a small business in Atlanta sells custom coffee mugs. For orders of 50 mugs or fewer, the price is ten dollars per mug. But if a customer orders more than 50, the price for the entire order drops to eight dollars per mug.
How do we write a function for the total cost?
Well, the first thing I notice is that the rule changes. There's a "break point" at 50 mugs. That's a huge clue that we need a piecewise function. It's a function with two different personalities.
For the first piece, when the number of mugs, let's call it x, is 50 or less, the cost is simple: ten times x.
For the second piece, when x is greater than 50, the cost is eight times x.
So our function, C of x, has two parts. It's ten x, if x is between one and fifty. And it's eight x, if x is greater than fifty.
Now, here's a common mistake I see every year. Students will create the function, but then they forget about the real-world context. The variable x represents mugs. You can't order negative five mugs, and you can't order half a mug. So we have to state a restriction: the domain, or the valid inputs for x, must be positive integers. Forgetting that is leaving the job half-done.
So, when you're working on these problems, remember you're wearing two hats. You're a mathematician choosing a function, and you're a scientist defining the rules of the world your function lives in. You can do this.
If you're modeling profit and there's an optimal price that maximizes it, a linear model can't capture that peak. It will incorrectly suggest that profit either increases or decreases forever.
Look at the data's shape. If it curves or has a clear high/low point, test a quadratic or other polynomial model. Use the finite differences test if you have a data table.
A function `h(t) = -16t^2 + 32t + 5` might be mathematically valid for `t = -10`, but `t` represents time in seconds after a ball is thrown. Negative time is meaningless in this context.
Always ask yourself: "What do my variables represent?" If it's items, they must be non-negative integers. If it's time, it must be non-negative. If it's a physical length, it must be positive.
Both functions curve upward, but they model fundamentally different types of growth. A quadratic model has a constant *second* difference (additive change), while an exponential model has a constant *ratio* (multiplicative change).
From a data table, check the differences first. If they aren't constant, then check the ratios of consecutive terms (`y2/y1`, `y3/y2`, etc.). If the ratios are constant, it's exponential. If the second differences are constant, it's quadratic.
In our mug example (`10x` if `x <= 50`, `8x` if `x > 50`), calculating the cost for exactly 50 mugs using the `8x` rule is incorrect. The definition explicitly includes 50 in the first rule.
Pay close attention to the inequality symbols (`<`, `>`, `≤`, `≥`). Carefully substitute your input value into the *correct* piece of the function based on its interval.
On the AP exam, articulating assumptions is a specific skill being tested (Skill 3.C). Simply providing a function without stating the conditions under which it is valid is an incomplete answer.
After you choose your model, take a moment to think: "What am I pretending is true for this model to work?" Write it down. (e.g., "Assuming no air resistance," or "Assuming the cost of materials remains constant.")