Chapter 14 — Front Matter
Purpose and Overview
This chapter closes the book. It does not introduce a new family of functions or a new algebraic technique. Instead, it brings together everything this textbook has developed — functions, transformations, trigonometry, analytic geometry, vectors, sequences, series, limits, and continuity — and asks a single overarching question: How do we use mathematics to describe, analyze, and explain the world?
That question is the heart of mathematical modeling. Every chapter in this book has been, in some sense, a modeling chapter. When you studied exponential functions, you were building tools for modeling growth and decay. When you studied trigonometric graphs, you were building tools for modeling periodic behavior. When you studied parametric equations, you were building tools for modeling motion. This chapter makes that theme explicit. It teaches you not just how to use individual models, but how to choose among them, how to interpret what they say, how to measure how well they work, and how to explain your reasoning in words.
This is also the chapter that prepares you most directly for Calculus I. Calculus is the study of change and accumulation. To use its tools well, you must know functions deeply — you must be able to look at a situation and say: this is the right function family; this is why; these are its parameters; and this is what it predicts. By the time you finish Chapter 14, precalculus should feel not like a list of disconnected topics, but like a coherent toolkit assembled over fourteen chapters for describing relationships, modeling behavior, and interpreting results.
Opening Question
A researcher studies three populations in a nature preserve. Population A grows rapidly at first and then levels off. Population B oscillates between high and low values on a roughly twelve-month cycle. Population C was introduced at a small initial value and appears to be doubling at a roughly constant rate. The researcher must identify the appropriate model for each population, build the model from data, predict future values, assess model fit, and explain the results to a non-specialist audience. She cannot do any of this by memorizing formulas. She needs to think about the structure of each situation. This chapter develops exactly those skills.
15Sections
10Objectives
7Capstone Problems
10+Worked Examples
Learning Objectives
- 1Explain the purpose and structure of mathematical modeling and describe the modeling cycle.
- 2Identify the appropriate function family for a real-world situation based on its behavior and structure.
- 3Build linear, quadratic, polynomial, rational, exponential, logarithmic, trigonometric, parametric, vector, and sequence models from descriptions, data, and contexts.
- 4Interpret the meaning of model parameters in real-world context.
- 5Compare multiple models for the same data and assess which fits best.
- 6Calculate and interpret residuals and model error.
- 7Use technology to support modeling, fitting, and interpretation without replacing mathematical reasoning.
- 8Write clear, accurate mathematical explanations that connect symbolic results to real-world meaning.
- 9Solve capstone problems requiring choosing a model, building it, interpreting it, and checking it.
- 10Demonstrate readiness for Calculus I by connecting function behavior to rates of change and accumulation.
Sections 14.1 – 14.2
What Modeling Is and How to Choose a Model
The Modeling Cycle · Function Families · Behavioral and Structural Evidence
Section 14.1
What Mathematical Modeling Does
A mathematical model is a mathematical structure — an equation, graph, table, sequence, or system — that represents a real-world situation in a simplified but useful way. The word model is important: a model is not the same as the real situation. It is a representation. A weather model does not contain the atmosphere; it represents the atmosphere using equations that capture enough relevant structure to make useful predictions.
Every mathematical model involves three essential steps. Represent: translate the situation into a mathematical form by choosing variables, identifying relationships, selecting a function family, and writing an equation or graph. Analyze: use the mathematical structure to answer questions — solve equations, find zeros, identify extrema, examine end behavior, or compute rates of change. Interpret: translate the result back into the language of the real situation. An answer of x = 7.3 means nothing without context, units, and explanation.
The Modeling Cycle — Six Stages
1. Identify the situation — What is being measured? What changes? What stays constant?
2. Represent mathematically — Choose variables. Assign units. Choose a function family. Write the equation.
3. Solve or analyze — Use algebra, trigonometry, graphing, or technology to extract information.
4. Interpret — Translate the result into the real context. Is the answer reasonable?
5. Validate — Compare predictions to known data. Where does the model break down?
6. Revise if needed — If the model does not fit well, refine the function family or parameters.
Modeling and the MGU Problem-Solving Method
The modeling cycle is the extended form of the MGU method: Read → Represent → Analyze → Solve → Check → Explain. Every problem in this chapter should be approached with all six stages, not just the algebra in the middle.
What Models Cannot Do
A model built from limited data should not be trusted far outside that data range — this is called extrapolation, and it must be treated cautiously. Estimating within the data range is called interpolation and is generally more reliable. A model also assumes the underlying relationship is stable. If conditions change, the model may no longer apply.
Section 14.2
Choosing a Function Family
Every modeling problem requires a central decision: which function family should represent this situation? This decision is guided by three kinds of evidence. Behavior of the data: does the output grow without bound, level off, oscillate, or grow faster and faster? Structure of the situation: is the rate of change constant, proportional to the current value, or cyclic? Shape of the graph: does it look like a line, parabola, wave, decay curve, or rational curve with asymptotes?
Function Family Guide
Linearf(x) = mx + b
- Output changes at a constant rate
- Graph is a straight line
- Use when: constant rate of change — hourly wages, simple depreciation, constant-speed distance
Quadraticf(x) = ax² + bx + c
- Output has a single turning point (maximum or minimum)
- Graph is a parabola
- Use when: projectile height, profit optimization, area with a fixed perimeter constraint
Exponentialf(t) = a · bᵗ or f(t) = a · eᵏᵗ
- Output grows or decays by a constant multiplicative factor per period
- Rate of change is proportional to the current value
- Use when: population growth, compound interest, radioactive decay, drug elimination
Logarithmicf(x) = a·log(x) + b
- Output grows quickly at first, then flattens dramatically
- Use when: intensity scales spanning many orders of magnitude — Richter, decibels, pH
Trigonometricf(t) = a·sin(bt + c) + d
- Output oscillates between a fixed maximum and minimum with a repeating period
- Use when: tides, temperature cycles, sound waves, electrical signals, hours of daylight
Rationalf(x) = p(x) / q(x)
- Output has restricted domain, asymptotes, and potentially complex local behavior
- Use when: inverse proportion, average cost functions, concentration models
Parametricx = f(t), y = g(t)
- Horizontal and vertical positions described by a third variable (usually time)
- Use when: two-dimensional motion, paths that fail the vertical line test
Vectorv = ⟨a, b⟩
- Describes both magnitude and direction
- Use when: velocity, force, displacement, navigation — direction matters
Sequenceaₙ = formula or recursive rule
- Defined only for positive integer inputs; discrete, not continuous
- Use when: monthly savings, population generations, compound interest periods
Student Warning — Choosing by Appearance Alone
Do not choose a function family just because the graph looks roughly curved. Check the structural evidence: constant differences (linear), constant ratios (exponential), constant second differences (quadratic), repeating cycles (trigonometric). A set of data points that looks curved might be exponential, quadratic, logarithmic, or power — only analysis of differences and ratios will tell you which.
Sections 14.3 – 14.6
Building Models by Function Family
Linear · Quadratic · Rational · Exponential · Logarithmic · Trigonometric · Parametric · Vector
Section 14.3
Linear, Quadratic, Polynomial, and Rational Models
Example 14.3.1
Building a Linear Model
- SituationA car rental company charges a flat fee of $40 plus $0.18 per mile driven. Write a model for the total cost C as a function of miles driven m. Find the cost for a 250-mile trip.
- ThinkThe cost increases at a constant rate of $0.18 per mile with a fixed base cost of $40. The rate of change is constant — this is a linear model.
- Represent
C(m) = 0.18m + 40
- Solve
C(250) = 0.18(250) + 40 = 45 + 40 = 85
- Check$45 for mileage plus $40 base = $85. Reasonable for 250 miles. ✓
Answer: C(m) = 0.18m + 40. A 250-mile trip costs $85.
Example 14.3.2
Building a Quadratic Model — Projectile Height
- SituationA ball is thrown upward from a height of 5 feet with an initial vertical velocity of 40 feet per second. Model the height h(t) and find the maximum height.
- ThinkProjectile motion under constant gravity follows the standard quadratic model: h(t) = −16t² + v₀t + h₀, where v₀ = 40 and h₀ = 5.
- Represent
h(t) = −16t² + 40t + 5
- SolveMaximum at the vertex. Using t = −b/(2a):
t = −40 / (2 · −16) = −40 / −32 = 1.25 seconds
h(1.25) = −16(1.5625) + 40(1.25) + 5 = −25 + 50 + 5 = 30 feet
- CheckBall starts at 5 ft, rises to 30 ft, then falls. Maximum at t = 1.25 s is plausible. ✓
Answer: h(t) = −16t² + 40t + 5. Maximum height is 30 feet at t = 1.25 seconds.
Example 14.3.3
Building a Rational Model — Average Cost
- SituationA factory's average cost per widget is A(x) = (5000 + 12x) / x, where x is number of widgets. What happens as production increases?
- ThinkRewrite to reveal structure: A(x) = 5000/x + 12. As x → ∞, the fraction 5000/x → 0.
- Analyze
A(x) = 5000/x + 12 → as x → ∞, A(x) → 12The horizontal asymptote y = 12 is the minimum achievable average cost.
- InterpretThe $5000 is a fixed cost (equipment, rent) spread over more units as production grows. The $12 is the per-unit variable cost. Greater production spreads the fixed cost more thinly, driving the average toward $12.
Answer: As production increases without bound, average cost approaches $12/widget. The asymptote represents the minimum achievable average cost.
Section 14.4
Exponential and Logarithmic Models
An exponential model applies when the rate of change of a quantity is proportional to the quantity itself — the larger the quantity, the faster it grows (or decays). This is the key structural test. The model takes the form f(t) = a · bᵗ (discrete periods) or f(t) = a · eᵏᵗ (continuous change), where a is the initial value, b is the growth factor, and k is the continuous rate.
Example 14.4.1
Exponential Growth — Bacterial Population
- SituationA bacterial culture begins with 200 bacteria and doubles every 3 hours. Write a model for P(t). Predict the population after 12 hours.
- ThinkDoubling every 3 hours means the growth factor is 2 per 3-hour period. The exponent must be t/3 to apply the doubling each period.
- Represent
P(t) = 200 · 2^(t/3)
- Solve
P(12) = 200 · 2^(12/3) = 200 · 2⁴ = 200 · 16 = 3200
- Checkt=0: 200 → t=3: 400 → t=6: 800 → t=9: 1600 → t=12: 3200. Each interval doubles. ✓
Answer: P(t) = 200 · 2^(t/3). After 12 hours, the population is 3200 bacteria.
Example 14.4.2
Exponential Decay — Half-Life
- SituationA radioactive substance has a half-life of 8 years. A sample begins with 500 grams. Write a model for A(t). How much remains after 24 years?
- Represent
A(t) = 500 · (1/2)^(t/8)
- Solve
A(24) = 500 · (1/2)^(24/8) = 500 · (1/2)³ = 500 · (1/8) = 62.5 grams
- Checkt=0: 500 → t=8: 250 → t=16: 125 → t=24: 62.5. Each period halves the amount. ✓
Answer: A(t) = 500 · (0.5)^(t/8). After 24 years, 62.5 grams remain.
Example 14.4.3
Logarithmic Scale — Richter Scale
- SituationTwo earthquakes measure 5.0 and 7.0 on the Richter scale. How many times more intense is the second?
- ThinkThe Richter scale is logarithmic base 10. Each whole-number increase represents a tenfold increase in measured amplitude.
- Solve
10^(7.0 − 5.0) = 10² = 100
- InterpretA magnitude-7.0 earthquake is 100 times more intense than a magnitude-5.0 earthquake. The logarithmic scale compresses an enormous range of physical intensities into a small range of numbers.
Answer: The magnitude-7.0 earthquake is 100 times more intense than the magnitude-5.0 earthquake.
Section 14.5
Trigonometric Models
Trigonometric models apply whenever the situation shows periodic behavior — behavior that repeats on a fixed cycle. The general model is:
General Trigonometric Model
f(t) = a · sin( (2π/T)(t − h) ) + k
a = amplitude: half the range from minimum to maximum
T = period: length of one complete cycle (in time units)
h = phase shift: horizontal offset of the cycle
k = midline: the average (center) value of the output
Example 14.5.1
Tidal Height Model
- SituationMaximum tidal height of 12 ft occurs at 4:00 a.m.; minimum of 2 ft near 10:00 a.m. Period is approximately 12.4 hours. Build a model H(t) where t is hours after midnight.
- Parameters
Amplitude: a = (12 − 2)/2 = 5 feet
Midline: k = (12 + 2)/2 = 7 feet
Period: T = 12.4 hours
Phase shift: h = 4 (maximum at t = 4)
- Model
H(t) = 5 cos( (2π/12.4)(t − 4) ) + 7
- CheckAt t=4: H = 5·cos(0) + 7 = 12 ✓ At t=10.2 (half period later): H = 5·cos(π) + 7 = 2 ✓
- InterpretAmplitude of 5 feet = variation from average. Midline of 7 feet = average tidal height. Phase shift of 4 places high tide at 4:00 a.m.
Answer: H(t) = 5 cos( (2π/12.4)(t − 4) ) + 7
Example 14.5.2
Seasonal Temperature Model
- SituationAverage monthly high temperature: 30°F in January (m=1), 90°F in July (m=7). Period is 12 months.
- Parameters
a = (90 − 30)/2 = 30 k = (90 + 30)/2 = 60 T = 12 h = 7
- Model
T(m) = 30 cos( (π/6)(m − 7) ) + 60
- CheckAt m=7: T = 30·cos(0) + 60 = 90 ✓ At m=1: T = 30·cos(−π) + 60 = −30 + 60 = 30 ✓
Answer: T(m) = 30 cos( (π/6)(m − 7) ) + 60
Section 14.6
Parametric and Vector Models
Example 14.6.1
Parametric Projectile Motion
- SituationA ball is launched from the origin with horizontal velocity 30 ft/s and vertical velocity 40 ft/s. Build a parametric model for its position.
- ThinkHorizontal and vertical motion are independent. Horizontal: no acceleration. Vertical: gravity at −16 ft/s².
- Model
x(t) = 30t
y(t) = −16t² + 40t
- InterpretAt any time t, the ball is at position (30t, −16t² + 40t). The time parameter preserves information about when the ball reaches each position — which is lost when you eliminate the parameter to write y as a function of x alone.
Answer: x(t) = 30t, y(t) = −16t² + 40t
Example 14.6.2
Vector Navigation Model
- SituationA ship travels with velocity vector ⟨12, 5⟩ knots (12 knots east, 5 knots north). Find speed and direction.
- Solve
Speed = |v| = √(12² + 5²) = √(144 + 25) = √169 = 13 knots
Direction: θ = arctan(5/12) ≈ 22.6° north of due east
- InterpretThe ship travels at 13 knots in a direction approximately 22.6° north of east. Both the speed (magnitude) and the heading (direction) are essential parts of the answer — neither alone is sufficient to describe the voyage.
Answer: Speed is 13 knots; direction is approximately 22.6° north of east.
Sections 14.7 – 14.12
Fitting Models, Interpreting Parameters, and Writing Explanations
Data Identification · Regression · Parameters · Residuals · Technology · Writing
Section 14.7
Fitting Models to Data
In many problems, you begin not with a description but with a table of measured values. You must first identify the appropriate function family, then estimate or calculate the parameters. Use these structural tests on equally spaced inputs:
| Function Family | Test | What to Check |
|---|
| Linear | First differences | Differences between consecutive outputs are approximately constant |
| Exponential | Ratios | Ratios of consecutive outputs are approximately constant |
| Quadratic | Second differences | Second differences (differences of differences) are approximately constant |
| Trigonometric | Cycle | Output rises and falls repeatedly with a fixed period |
| Logarithmic | Rate of growth | Output grows rapidly at first, then very slowly |
Example 14.7.1
Identifying Exponential Data
- Data
t: 0 1 2 3 4
P: 100 150 225 337.5 506.25
- TestCheck ratios: 150/100 = 1.5 · 225/150 = 1.5 · 337.5/225 = 1.5 · 506.25/337.5 = 1.5. Constant ratio confirms exponential.
- Model
P(t) = 100 · 1.5^t
- CheckAt t=0: 100 ✓ At t=4: 100 · 5.0625 = 506.25 ✓
Answer: P(t) = 100 · 1.5^t. The colony grows by a factor of 1.5 per hour.
Technology Note — Regression Types
Common regression types on graphing calculators: LinReg (y = ax + b), QuadReg (y = ax² + bx + c), ExpReg (y = a·bˣ), LnReg (y = a·ln(x) + b), SinReg (y = a·sin(bx+c)+d), PwrReg (y = axᵇ). Always enter data, plot a scatter, then choose the appropriate regression. Record the equation and the r² value. Graph the regression curve over the scatter plot to check visually. For SinReg, confirm the calculator is in radian mode.
Section 14.8
Interpreting Parameters
Every parameter in a model carries real-world meaning. Treating parameters as abstract numbers to be solved for and then forgotten misses half the value of modeling. A model is only useful when you can explain what it says about the real situation.
Example 14.8.1
Interpreting Exponential Parameters — Medication Decay
- Model
A(t) = 800 · e^(−0.14t) where t is in hours and A is in milligrams
- Interpret aa = 800 mg: The initial dose at t = 0 (moment of administration) is 800 milligrams.
- Interpret kk = −0.14: The medication decays continuously at a rate of 14% per hour. The negative sign confirms decay rather than growth.
- Half-lifeSet A(t) = 400 and solve:
0.5 = e^(−0.14t)
ln(0.5) = −0.14t
t = ln(0.5) / (−0.14) ≈ 4.95 hours
- InterpretEvery 4.95 hours, the amount remaining is cut in half.
Answer: Initial dose 800 mg; 14% per hour decay rate; half-life approximately 4.95 hours.
Example 14.8.2
Interpreting Trigonometric Parameters — AC Circuit
- Model
I(t) = 3.5 sin(120πt) where t is in seconds, I is in amperes
- Amplitudea = 3.5 amperes. The current varies between −3.5 and +3.5 amperes.
- Period
T = 2π / (120π) = 1/60 seconds → frequency of 60 Hz. This is the standard North American AC power frequency.
- InterpretThe circuit carries alternating current with a peak of 3.5 A, completing 60 full oscillations per second.
Answer: Amplitude 3.5 A, period 1/60 s, frequency 60 Hz — standard AC power.
Section 14.9
Comparing Models
When more than one function family appears to fit a data set, compare models by examining: (1) how well each fits the data (residuals and r²), (2) whether the model makes physical sense outside the data range, and (3) whether the model is parsimonious — a simpler model is preferred when two models fit equally well.
The coefficient of determination r² measures how much of the variation in the output is explained by the model. An r² close to 1 indicates a good fit. But r² alone is not sufficient: check residuals and physical plausibility.
Example 14.9.1
Comparing Linear and Exponential Models
- Data
t: 0 5 10 15 20
P: 1000 1500 2200 3300 5000
- LinearP(t) = 202t + 800, r² = 0.96
- ExponentialP(t) = 1000 · e^(0.082t), r² = 0.999
- CompareThe linear model has high r² but does not match the increasing growth rate visible in the data. The differences between consecutive outputs are not constant; the ratios are closer to constant. The exponential model has higher r² and matches the structural behavior — the rate of growth increases proportionally to the population.
Answer: The exponential model is the better choice — higher r² and structurally correct for a growing population.
Section 14.10
Model Error and Residuals
A residual is the difference between an actual data value and the model's predicted value at the same input: Residual = Actual − Predicted. A positive residual means the model underestimates; a negative residual means the model overestimates.
A good model has residuals that are small in size and randomly distributed around zero. If residuals form a curve or systematic pattern, the wrong function family was chosen.
Example 14.10.1
Computing and Interpreting Residuals
- ModelLinear model P(t) = 202t + 800 applied to the population data from Example 14.9.1.
- Residuals
| t | Actual | Predicted | Residual |
|---|
| 0 | 1000 | 800 | +200 |
| 5 | 1500 | 1810 | −310 |
| 10 | 2200 | 2820 | −620 |
| 15 | 3300 | 3830 | −530 |
| 20 | 5000 | 4840 | +160 |
- InterpretThe residuals follow a clear pattern — negative in the middle, positive at the ends — not random scatter. This systematic shape reveals that the linear model is structurally wrong for this data. An exponential or other nonlinear model should be tried.
Answer: Systematic residual pattern indicates the linear model is wrong. A nonlinear (e.g., exponential) model should be used instead.
Mean Absolute Error
MAE = (1/n) · Σ|Residual_i|
A smaller MAE indicates a better fit. MAE is in the same units as the output variable, making it easy to interpret in context.
Section 14.11
Technology-Assisted Modeling
Technology can plot scatter plots, perform regression, evaluate models quickly, plot residuals, and graph functions with precision. Technology cannot decide which function family is appropriate, interpret what parameters mean in context, or determine whether a model makes physical sense. A student who enters data into a calculator, receives an equation, and writes that equation as the answer has completed only part of the modeling process.
Graphing Calculator Modeling Workflow
Step 1: Enter data into two lists (inputs and outputs).
Step 2: Plot a scatter plot. Observe the shape. Choose a function family.
Step 3: Run the appropriate regression from the statistics menu.
Step 4: Record the equation and the r² value. Graph over the scatter plot.
Step 5: Examine the residual plot. Random scatter → good fit. Curved pattern → wrong family.
Step 6: Use the model to answer the question. Interpret the result in context.
Technology Note — Window Settings
Set the graphing window to reflect the meaningful domain of the real situation. A tidal model over 24 hours should show x from 0 to 24. A population model over 50 years should show x from 0 to 50. Do not use a default window that obscures the real behavior of interest.
Section 14.12
Writing Mathematical Explanations
Mathematics is a language, and using it well includes explaining results in words. Engineers explain calculations to clients. Scientists communicate model results in papers. Financial analysts describe predictions to stakeholders. In every case, the explanation must connect the mathematical result to its real-world meaning. A number without context is not an answer.
Structure of a Good Mathematical Explanation
1. State what was done. Name the model or method used.
2. State the result. Give the equation, value, or conclusion.
3. Interpret the result. Explain what it means in the real situation.
4. Address limitations. Acknowledge assumptions and where the model may not apply.
Example 14.12.1
Writing an Explanation for a Non-Specialist Audience
- SituationWrite a clear explanation of the model A(t) = 800 · e^(−0.14t) for a nurse who is not a mathematician.
- Draft"This model predicts how much medication remains in the patient's bloodstream at any time t hours after the dose is given. At the moment of administration, the amount is 800 milligrams. The medication is eliminated continuously, decreasing by approximately 14 percent each hour. Based on this model, the amount falls to about 400 milligrams after roughly 5 hours. After 10 hours, approximately 200 milligrams remain. This model assumes the rate of elimination is constant over time. For accurate clinical decisions, this model should be verified against measured blood concentration values."
- Why It WorksNames the model (exponential) · explains the initial value (800 mg) · interprets the rate (14%/hr) · gives specific predictions (half-life ~5 hrs) · acknowledges the constant-rate assumption.
Common Writing Mistakes
Stating the equation without interpretation. The equation alone is not an explanation. · Using undefined variables. Always say what each variable represents. · Omitting units. "The result is 62.5" is incomplete; "62.5 grams remaining after 24 years" is complete. · Claiming more accuracy than the model provides. If the model is based on limited data, do not claim precise long-range predictions.
Sections 14.13 – 14.14
Capstone Problems and Common Modeling Mistakes
Finance · Ecology · Temperature · Physics · Navigation · Sound · Model Comparison · Eight Recurring Errors
Section 14.13
Capstone Modeling Problems
Each capstone problem requires choosing an appropriate model, building it, interpreting parameters, assessing fit, and explaining results. Approach every problem using the full modeling cycle: Read → Represent → Analyze → Solve → Check → Explain.
Capstone Problem 1
Finance — Continuous Growth
- An investor places $5000 in an account earning 6% interest compounded continuously. Build a model for the account balance A(t) where t is years.
- Find the balance after 10 years.
- Find the time required for the investment to double.
- Interpret the growth rate k = 0.06 in plain language.
Capstone Problem 2
Ecology — Exponential Population Growth
- A ranger monitors deer: Year 0 = 120, Year 1 = 144, Year 2 = 173, Year 3 = 207, Year 4 = 249, Year 5 = 299. Examine the ratios of consecutive values. What function family is suggested?
- Build an exponential model for the population.
- Predict the population in Year 8.
- Compute the residuals for Years 1 through 5. Is the model appropriate? Explain.
Capstone Problem 3
Seasonal Temperature
- Monthly high temperatures (°F), m = 1 for January: 18, 22, 35, 48, 60, 70, 75, 72, 60, 46, 32, 22. Identify the maximum and minimum values and their months.
- Compute the amplitude, midline, and period.
- Build a trigonometric model T(m).
- Predict the temperature in month 9 and compare to the actual value.
- What does the midline represent physically?
Capstone Problem 4
Physics and Parametric Motion
- A soccer ball is kicked with horizontal velocity 25 ft/s and vertical velocity 35 ft/s from the ground. Build parametric equations for x(t) and y(t).
- Find the time when the ball reaches its maximum height.
- Find the maximum height.
- Find the horizontal distance when the ball lands.
- Eliminate the parameter t to write y as a function of x.
Capstone Problem 5
Navigation and Vectors
- A drone has velocity ⟨20, 15⟩ ft/s. Wind adds ⟨−5, 8⟩ ft/s. Find the resultant velocity vector.
- Find the drone's actual speed.
- Find the direction of travel as an angle north of east.
- After 30 seconds, what are the drone's coordinates from its starting point?
Capstone Problem 6
Sound Wave Modeling
- A tuning fork model: P(t) = 0.002 sin(880πt), where P is air pressure in pascals and t is in seconds. Identify the amplitude and period.
- Find the frequency in Hz.
- Musical note A above middle C has frequency 440 Hz. What note does this model represent? (Hint: Compare frequencies and recall that doubling the frequency raises the pitch by one octave.)
- Describe one full cycle of this wave in words.
Capstone Problem 7
Comparing Three Models
- Three models for a population data set: Model A (linear): P(t) = 300t + 1000, r² = 0.91 · Model B (quadratic): P(t) = 5t² + 200t + 1000, r² = 0.97 · Model C (exponential): P(t) = 1000·e^(0.21t), r² = 0.99. Based on r² alone, which appears best?
- Describe the long-run behavior (end behavior) of each model as t → ∞.
- Which model is most physically reasonable for a growing population? Explain your reasoning.
- If the data shows consecutive population values with nearly constant ratios, which model is confirmed?
Section 14.14
Common Modeling Mistakes
Mistake 1 — Choosing by appearance rather than structure
A set of curved data is not necessarily quadratic. Check constant differences (linear), constant ratios (exponential), or constant second differences (quadratic) before committing to a function family. The visual shape of the graph is only one piece of evidence.
Mistake 2 — Confusing exponential growth with linear growth
P(t) = 300t + 1000 for a population that doubles at each step
Linear growth adds a constant amount per period. Exponential growth multiplies by a constant factor. When the rate of growth itself increases over time, a linear model fails — often dramatically — at later time values.
Mistake 3 — Ignoring units
Every parameter carries units. The slope of a linear model has units of output per input. The period of a trigonometric model has units of time. The decay constant has units of per-time. Unitless parameters produce nonsensical interpretations.
Mistake 4 — Trusting extrapolation
A model valid for 20 years of data may dramatically overpredict at year 100. Trigonometric models assume the period and amplitude remain stable indefinitely. Always state the domain over which a model is supported by data.
Mistake 5 — Treating a high r² as proof the model is correct
r² measures fit to the given data. It does not guarantee the most appropriate function family, good performance on new data, or physical plausibility. Always also check residuals and ask whether the model makes sense outside the data range.
Mistake 6 — Forgetting to interpret parameters
P(t) = 1000 · e^(0.15t) [written as the final answer with no explanation]
The equation must be interpreted: the initial population is 1000, and the population grows continuously at a rate of 15% per year. Without interpretation, the model is just symbols.
Mistake 7 — Incorrect trigonometric model parameters
Period = b (from f(t) = a·sin(bt) + d)
The period is 2π/b, not b. Amplitude is |a|. Phase shift from f(t) = a·sin(b(t − h)) + k is h (not bh). Midline is k. Each parameter requires a careful formula.
Mistake 8 — Describing a model without connecting it to the real situation
The equation is part of the answer, never the whole answer. Every model must be described in terms of what the variables represent, what the parameters mean in context, and what the model predicts about the real situation. A number without context is incomplete.
Chapter 14 — Back Matter
Chapter Summary
- Sec 14.1A mathematical model represents a real situation using a mathematical structure. Every model involves three steps: represent, analyze, and interpret. The modeling cycle has six stages. Extrapolation beyond the data range is unreliable.
- Sec 14.2Choosing a function family requires evidence from behavioral patterns, structural situation, and graph shape. Each function family has a characteristic behavior: constant-rate (linear), turning-point (quadratic), multiplicative (exponential), scale-based (logarithmic), oscillatory (trigonometric), directed motion (parametric/vector), discrete (sequence).
- Sec 14.3Linear models use constant rate of change and initial value. Quadratic models use the vertex for max or min problems. Rational models describe average cost, inverse proportion, and asymptotic long-run behavior.
- Sec 14.4Exponential models apply when the rate of change is proportional to the current value. The base or continuous rate determines growth or decay. Logarithmic models apply to scale and the inverse of exponential relationships.
- Sec 14.5Trigonometric models apply to periodic behavior. The four key parameters — amplitude, period, phase shift, and midline — each carry real-world meaning. Cosine is preferred when the maximum occurs at a known time.
- Sec 14.6Parametric models describe two-dimensional motion by separating x and y as functions of time. Vector models describe magnitude and direction simultaneously. Both are essential for physics, engineering, and navigation.
- Sec 14.7Identifying function families from data: constant differences → linear; constant ratios → exponential; constant second differences → quadratic. Technology supports regression, but choosing the family remains a human mathematical decision.
- Sec 14.8Every parameter carries real-world meaning. Initial value, growth rate, amplitude, midline, half-life, and period must all be interpreted in the context of the real situation. Writing the equation without interpretation is incomplete.
- Sec 14.9r² measures how much variation in the data is explained by the model. Higher r² is better, but r² alone is not sufficient — check physical plausibility and residual patterns as well. Simpler models are preferred when fits are comparable.
- Sec 14.10Residual = Actual − Predicted. A residual plot that shows random scatter confirms the model is appropriate. A curved or systematic residual pattern reveals that the wrong function family was chosen.
- Sec 14.11Technology performs regression, evaluates, and plots. Technology cannot choose the function family, interpret parameters, or determine physical plausibility. The regression equation must be interpreted in context.
- Sec 14.12A good mathematical explanation states the method, gives the result, interprets it in context, and acknowledges limitations. Defining variables, including units, and addressing model assumptions are required parts of any complete explanation.
- Sec 14.13Capstone problems draw on all function families and require the full modeling cycle: identify, represent, analyze, solve, check, interpret, and explain. Finance, ecology, temperature, motion, navigation, sound, and model comparison are all modeled.
- Sec 14.14Eight recurring modeling errors: choosing by appearance; confusing linear and exponential; ignoring units; trusting extrapolation; relying on r² alone; forgetting parameter interpretation; incorrect trig parameters; omitting real-world context from the explanation.
- Sec 14.15Calculus builds on precalculus fluency with functions, graphs, limits, rates of change, and modeling. Every function family in this book reappears in Calculus I. The derivative formalizes average rate of change; the integral formalizes accumulation; limits underlie both.
Key Terms Review
mathematical model
A mathematical structure — equation, graph, table, or sequence — that represents a real-world situation in a simplified but useful way.
function family
A category of functions sharing the same general form and type of behavior — linear, quadratic, exponential, trigonometric, etc.
parameter
A constant in a model equation that carries real-world meaning — such as initial value, growth rate, amplitude, or period.
residual
The difference between an actual data value and the model's predicted value: Residual = Actual − Predicted.
r² (coefficient of determination)
A measure of how much of the variation in the output data is explained by the model. Ranges from 0 to 1; closer to 1 indicates a better fit.
regression
A method of finding the curve of best fit for a set of data by minimizing the total error between predicted and actual values.
extrapolation
Using a model to predict values outside the range of the data it was built from. Generally less reliable than interpolation.
interpolation
Using a model to estimate values within the range of the data. Generally more reliable than extrapolation.
mean absolute error
The average of the absolute values of all residuals. Measures average model error in the same units as the output variable.
amplitude
In a trigonometric model, half the distance between the maximum and minimum values. Controls the vertical scale of the wave.
period
The length of one complete cycle in a trigonometric model. Equal to 2π/b in the form a·sin(bt + c) + d.
phase shift
The horizontal offset of a trigonometric model from its standard starting position. Equal to h in a·sin( (2π/T)(t − h) ) + k.
midline
The horizontal line halfway between the maximum and minimum of a trigonometric model. Equal to the vertical shift k.
half-life
In an exponential decay model, the time required for the quantity to fall to half of its current value.
modeling cycle
The six-stage process of mathematical modeling: identify → represent → solve → interpret → validate → revise.
calculus readiness
Fluency with functions, graphs, transformations, limits, rates of change, and modeling that prepares a student to succeed in Calculus I.
Concept Review Questions
- What is a mathematical model, and what three essential steps does every model involve?
- Describe the modeling cycle. What are the six main stages?
- What behavioral features suggest a linear model? An exponential model? A trigonometric model?
- Explain the difference between extrapolation and interpolation, and why extrapolation is less reliable.
- How do you identify an exponential model from a data table of equally spaced inputs?
- What does a residual measure? What does a curved residual plot indicate about the chosen model?
- What does r² measure? Is a high r² value by itself sufficient to confirm that a model is appropriate? Explain.
- In the trigonometric model f(t) = a·sin( (2π/T)(t − h) ) + k, what does each parameter represent in a real-world context?
- In an exponential model f(t) = a·e^(kt), what does a negative value of k indicate?
- Why is writing a mathematical explanation of a model an essential part of the modeling process?
- Name three ways in which precalculus directly prepares students for Calculus I.
- What is the relationship between the average rate of change in precalculus and the derivative in calculus?
Function Family Review
For each description, name the most appropriate function family and write a general model form.
- The temperature of a cooling liquid decreases by a fixed percentage of its current value per minute.
- The height of a projectile increases, reaches a peak, and then falls.
- The amount of daylight per day varies on an annual cycle.
- The average cost of production decreases as more units are produced but never falls below a fixed minimum.
- A city's population grows by the same fixed number of people per year.
- Sound intensity perceived decreases dramatically as distance increases, following a scale spanning many orders of magnitude.
- The voltage in an AC circuit oscillates between positive and negative values 60 times per second.
- A ball's horizontal and vertical positions must be tracked separately over time.
- A savings account balance multiplies by a constant factor each year.
- A drone's motion requires describing both its speed and the direction it travels.
Model Selection Practice
For each data table, examine differences or ratios to identify the function family, then build the model.
Examine first differences. Identify the function family. Build the model.
Examine ratios. Identify the function family. Build the model.
Examine differences and second differences. Identify the function family. Build the model.
A data set shows tidal heights measured every 6 hours over 48 hours, oscillating between 2 and 14 feet. What is the most appropriate function family? Write a general model structure with identified parameters.
Data Modeling Problems
- A sample of carbon-14 has an initial mass of 50 grams and a half-life of 5730 years.
(a) Build an exponential model for the remaining mass A(t).
(b) How much remains after 11,460 years?
(c) In how many years will only 10 grams remain?
- A spring causes a mass to oscillate with maximum displacement 8 cm above rest, minimum 8 cm below, and period 4 seconds. At t = 0 the mass is at maximum displacement.
(a) Write a trigonometric model for d(t).
(b) Find the displacement at t = 1 second.
(c) Find the first time the mass returns to rest.
- Drug concentration in the bloodstream: t = 0 → 200 mg/L; t = 2 → 150; t = 4 → 112.5; t = 6 → 84.4; t = 8 → 63.3.
(a) Check ratios to confirm an exponential model.
(b) Build the exponential model.
(c) Predict the concentration after 12 hours.
(d) Interpret the decay factor in context.
Error Analysis
Each problem shows a student error. Identify it and write the correct answer.
Problem 1 — Amplitude Error
Maximum = 20, minimum = 4 → amplitude = 20 − 4 = 16
What is the error? What is the correct amplitude? Write the corrected model form.
Problem 2 — Misidentifying Function Family
Ratios: 8/4 = 2, 14/8 = 1.75, 22/14 = 1.57 → "approximately constant at 1.8, so exponential"
The ratios are not constant — they are decreasing. What does this tell you about the function family? What should the student do instead?
Problem 3 — Misreading Period
In T(m) = 25 cos(π/6 · (m − 7)) + 55, the student writes: "Period = π/6"
Identify the error. What is the correct period? What is the formula relating b and T?
Problem 4 — Wrong Model Type
Data points (0, 500) and (10, 250) → "f(t) = 500 − 25t"
A classmate objects to the linear model. Who is correct, and why? What model should be used? Show how to build it from the two data points.
Chapter 14 Checkpoint
Chapter 14 Checkpoint
Complete without assistance — assesses all key modeling, interpretation, and calculus readiness skills from Chapter 14
Part A — Model Selection
- The number of active users grows by approximately 12% per month. Name the function family and justify your choice.
- A Ferris wheel car oscillates between 5 and 85 feet over a 3-minute cycle. Name the family and identify all parameters.
- Average production cost per unit decreases as units increase, approaching a fixed lower limit. Name the family.
- A ball has both horizontal and vertical velocity; both positions must be tracked separately. Name the modeling approach.
Part B — Building Models
- A city's population was 50,000 in 2010 and grew to 72,500 in 2020. Assuming exponential growth, write a model P(t) where t is years after 2010. Predict the population in 2030.
- A city's average temperature is 35°F in January (m=1) and 85°F in July (m=7), with a 12-month sinusoidal cycle. Build T(m).
- A ball is kicked with horizontal velocity 20 ft/s and vertical velocity 32 ft/s. Write parametric equations. Find the maximum height and horizontal distance when it lands.
Part C — Interpreting and Comparing
- Model: C(t) = 400 · e^(−0.2t), t in hours.
(a) What is the initial concentration?
(b) What does −0.2 represent?
(c) Find the half-life.
(d) Write a one-paragraph explanation for a patient.
- Two models for a population: Model A: P(t) = 120t + 1000, r² = 0.93; Model B: P(t) = 1000·1.12^t, r² = 0.99. Compare both models. Which is more appropriate, and why? Address both statistical fit and physical plausibility.
Part D — Calculus Readiness
- Explain in your own words the connection between the average rate of change in precalculus and the derivative in Calculus I.
- Name three function families studied in this book and describe one key property that makes each important in Calculus I.
- A function is defined by f(x) = 3·e^(0.5x). Without computing a derivative, explain what you expect about the rate of change as x increases. Is it increasing, decreasing, or constant? Justify your answer from the structure of the function.
Looking Ahead — Calculus I and Beyond
The Gateway Is Open
In Calculus I, you will formalize the limit concept introduced in Chapter 13, define the derivative as the limit of the difference quotient, and compute derivatives of polynomial, trigonometric, exponential, and logarithmic functions using differentiation rules. Integration will be introduced as the inverse of differentiation and as a tool for computing area and accumulation.
Every function family you have studied reappears. Polynomial terms are differentiated by the power rule. Trigonometric functions produce cyclic derivatives: d/dx(sin x) = cos x. The exponential function eˣ is its own derivative. Parametric equations enable motion problems and arc length. Sequences and series lead to infinite series in Calculus II. Vectors and parametric forms extend into multivariable calculus.
Beyond Calculus I: differential equations, where exponential and trigonometric models from this book describe oscillation, growth, and physical systems. Linear algebra, with roots in the vector work of Chapter 12. Probability and statistics, building on the modeling skills of Chapter 14. And every applied discipline — physics, engineering, finance, data science, computer graphics, robotics, and scientific computing — that speaks mathematics as its working language.
MGU Library Connections
- Calculus I — Chapters 1–5: Limits, Derivatives, and Integration
- Appendix R: Calculus I Readiness Check — full textbook back matter
- Appendix G: Trigonometric Identities Reference
- Appendix F: Unit Circle Reference
- Appendix K: Exponential and Logarithmic Rules Reference
- Appendix P: Limit and Continuity Preview Reference
- MGU Applied Mathematics Series: Physics and Engineering Applications
- MGU Finance and Economics Series: Financial Mathematics
- MGU Computing and Data Science Series: Data Science Foundations
- MGU Dictionary entries: model, parameter, residual, regression, derivative, integral, limit