9.3 The Mathematical Modeling Cycle & Secondary Applications

9.3 The Mathematical Modeling Cycle & Secondary Applications

Mathematical modeling is the bridge connecting abstract mathematical structures to the complexities, ambiguities, and dynamic systems of the physical, biological, and social worlds. Codified in the Common Core State Standards as Standard for Mathematical Practice 4 ("Model with mathematics") and extensively elaborated in the landmark GAIMME Report (Guidelines for Assessment & Instruction in Mathematical Modeling Education by SIAM and COMAP), modeling is not an isolated instructional topic—it is an authentic habit of mind. For NBPTS AYA Mathematics candidates, orchestrating authentic modeling experiences requires moving secondary students beyond rote, pre-packaged textbook word problems toward genuine inquiry where students define variables, formulate assumptions, select appropriate functional representations, and critically evaluate model limitations (NBPTS Standards II, III, and VI).

---

1. The Mathematical Modeling Cycle: Framework & Stages

The mathematical modeling cycle is an iterative, non-linear process. Rather than proceeding as a rigid procedural checklist, modeling requires students to navigate continuously between the "real world" and the "mathematical world," looping back to refine models whenever predictions fail empirical validation.

+-----------------------------------------------------------------------------+
|                     THE MATHEMATICAL MODELING CYCLE                         |
|                                                                             |
|   [REAL WORLD]                                      [MATHEMATICAL WORLD]    |
|                                                                             |
|   1. Messy Problem  ------(Identify & Simplify)----->  2. Assumptions &     |
|      Context                                              Variables         |
|         ↑                                                     │             |
|         │                                                 (Formulate)       |
|    (Refine /                                                  ↓             |
|     Iterate)                                           3. Mathematical      |
|         │                                                 Model             |
|         │                                                     │             |
|   6. Validating &   <-----(Interpret & Contextualize)-- 4. Computing        |
|      Limitations                                          Solutions         |
|         │                                                                   |
|         ↓                                                                   |
|   5. Real-World Reporting & Implementation                                  |
+-----------------------------------------------------------------------------+

The Seven Operational Stages of Modeling

1. Identifying the Problem: Encountering an open-ended, messy real-world situation and distilling a tractable mathematical question. (e.g., "How much water is wasted annually by leaking plumbing fixtures in our municipal school district?")
2. Making Assumptions & Defining Variables: Identifying essential factors and explicitly disregarding negligible noise. Defining parameters, independent and dependent variables, units of measure, and operational constraints.
3. Formulating the Model: Translating real-world relationships and assumptions into mathematical language—geometric representations, systems of equations, differential equations, recurrence relations, or statistical distributions.
4. Computing Solutions: Executing mathematical and computational operations—algebraic manipulation, calculus optimization, matrix operations, statistical regression, or computational dynamic simulations—to generate mathematical results.
5. Interpreting Results: Translating mathematical solutions back into the original real-world context, analyzing practical significance, and attaching real-world units.
6. Validating Conclusions & Limitations: Comparing model predictions against observed empirical data, evaluating reasonableness, conducting sensitivity analysis, and testing extreme boundary conditions (t → 0, t → ∞).
7. Iterating and Reporting: If the model fails validation or exhibits unacceptable error margins, returning to Stage 2 or 3 to adjust assumptions or select an alternate model family. Communicating findings clearly to stakeholders with transparent discussion of limitations.

---

2. Open-Ended Modeling vs. Routine Word Problems

Secondary mathematics curricula have historically conflated routine, pseudo-contextual word problems with genuine mathematical modeling. Accomplished educators maintain sharp distinctions between these instructional modes:

[!NOTE] Classroom Contrast Example:

• Routine Word Problem: "A water tank shaped like a cylinder with radius 4 m and height 10 m drains at 3 m³/min. How long does it take to empty?"
• Authentic Modeling Task: "A municipal water tower must serve a community of 15,000 residents. During a power outage where supply pumps shut down, design the dimensions of the elevated tank to ensure adequate emergency water reserve and water pressure for at least 36 hours."

---

3. Selecting Model Families Based on Empirical and Structural Phenomena

A critical modeling competency is selecting an appropriate mathematical function family based on the intrinsic rate of change and physical characteristics of the observed phenomenon.

Comparison of Common Secondary Modeling Families

Detailed Mathematical Mechanics of Model Selection

1. Quadratic Models: Constant Acceleration and Revenue Maximization

Quadratic functions govern systems where the rate of change itself changes at a constant rate (d²y/dt² = constant). In physics, uniform gravitational acceleration g produces the vertical trajectory equation h(t) = -½gt² + v₀ t + h₀. In economics, if the demand for a product decreases linearly with price (q(p) = a - bp), total revenue is inherently quadratic: R(p) = p · q(p) = p(a - bp) = -bp² + ap Maximizing revenue requires finding the vertex p* = -a / (2(-b)) = a / (2b).

2. Exponential Growth vs. Decay

Exponential models arise when the rate of change is directly proportional to the current state: dy/dt = ky ⟹ y(t) = y₀ e^(kt)

• If k > 0, the system experiences exponential growth (e.g., uninhibited bacterial proliferation N(t) = N₀ 2^(t/d)). In discrete time, successive ratios yₙ₊₁ / yₙ = b are constant.
• If k < 0, the system experiences exponential decay. Radioactive isotopes decay according to N(t) = N₀ e^(-λt), where the half-life is t₁/₂ = (ln 2) / λ. Newton's Law of Cooling models thermal equilibration as T(t) = T_env + (T₀ - T_env)e^(-kt).

---

4. The Logistic Growth Model & Dynamic Carrying Capacities

While exponential growth provides a reasonable model for early stages of biological proliferation or disease spread, physical resources (space, food, hosts) are finite. Unchecked exponential growth predicts infinite quantities (y → ∞), which violates physical reality. The logistic growth model resolves this limitation by introducing a carrying capacity K.

Analytical Structure of the Logistic Model

The logistic function is the unique solution to the non-linear Verhulst differential equation: dP/dt = r P (1 - P / K) where r is the intrinsic growth rate and K is the maximum sustainable carrying capacity.

Solving this separable differential equation yields the explicit closed-form function: P(t) = K / (1 + A e^(-rt)), where A = (K - P₀) / P₀

  P(t) ^
       |                                    --- Asymptote P = K (Carrying Capacity)
     K |...........................--------------------
       |                         .´
       |                       .´
   K/2 |----------------------* (Inflection Point: Maximum Growth Rate dP/dt)
       |                    .´
       |                 .´
    P₀ |__________.----´
       +---------------------------------------------> t (Time)

Core Properties of the Logistic S-Curve

1. Early Stage (P ≪ K): When the population is small relative to carrying capacity, the factor (1 - P/K) ≈ 1, and the model behaves as pure exponential growth: dP/dt ≈ rP.
2. Inflection Point (P = K/2): Differentiating the growth rate with respect to P shows that maximum absolute growth occurs precisely when the population reaches half of the carrying capacity: d/dP [rP - (r/K)P²] = r - 2(r/K)P = 0 ⟹ P = K / 2 At P = K/2, the second derivative d²P/dt² = 0, representing the transition from concave up (accelerating growth) to concave down (decelerating growth).
3. Asymptotic Convergence (t → ∞): As time increases, e^(-rt) → 0, forcing lim_{t → ∞} P(t) = K. The carrying capacity serves as a horizontal asymptote.

---

5. Secondary Classroom Orchestration & Rubric-Based Assessment

Promoting Student Autonomy in Assumption-Making

A foundational challenge in orchestrating modeling tasks is resisting the temptation to resolve ambiguity for students. When students confront an open modeling task, they frequently ask: "Do we need to account for sales tax?" or "Should we include leap years?"

• Accomplished Teacher Stance: Rather than answering directly, the teacher turns the question back to the student: "What impact would including sales tax have on your final decision? If it changes the outcome by less than 1%, is it worth the additional mathematical complexity?"
• The Art of Modeling: Modelers operate under the maxim formulated by statistician George Box: "All models are wrong, but some are useful." Students must learn that assumptions are deliberate trade-offs between realism and mathematical tractability.

Scaffolding Without Lowering Cognitive Demand

Accomplished educators provide structural scaffolds that maintain rigor:

• Fermi Estimation Prompts: Guiding students to establish order-of-magnitude bounds using powers of ten before constructing complex functional models.
• Variable Matrix Worksheets: Requiring student teams to categorize variables into Controlled Parameters, Independent Variables, and Dependent Variables with explicit units.
• Extreme Case Analysis: Instructing students to test what their mathematical function predicts when x = 0, x → ∞, or when parameters are negative. If a model predicts a negative population or infinite speed, the model's domain boundaries or formulation must be revised.

Multi-Dimensional Assessment Rubrics for Modeling

Grading modeling tasks strictly on whether students obtained a specific numerical answer destroys the authenticity of modeling. Assessment must evaluate reasoning across the entire GAIMME modeling cycle:

1. Framing & Assumptions (25%): Clearly articulates the real-world problem; identifies necessary variables; states simplifying assumptions with logical justifications.
2. Mathematical Formulation & Execution (25%): Selects an appropriate model family; executes algebraic, calculus, or numerical computations accurately; uses appropriate technology.
3. Interpretation & Real-World Synthesis (25%): Explains mathematical outputs in context with correct units; directly addresses the primary real-world question.
4. Validation, Sensitivity & Iteration (25%): Tests model predictions against empirical data or physical limits; analyzes how sensitive outputs are to changes in assumptions; articulates specific model limitations and proposes iterations.