Skip to content
Tutorials Center
My learning

Topics

Lesson 6 of 46 · The Modelling Cycle

The Modelling Cycle: Overview

Chapter 2 overview: the six-step modelling cycle from defining the question to reporting results, why it loops, and a first Fermi estimate you can run.

Experienced modellers do not start by writing equations. They start by asking, writing, doubting and sketching, and only then calculate. That order is not a matter of taste: it is what separates a model that answers a question from a pile of correct but useless algebra.

This chapter teaches the workflow professionals use, and you will apply it to one running example, a cooling cup of coffee, so you can see every step on the same problem.

Six numbered steps arranged in a loop: Define, Assume, Build, Solve, Check, Report
The modelling cycle. After the sixth step you usually return to the first with a sharper question.

The big picture

The modelling cycle has six steps. Different textbooks and syllabi name them slightly differently (formulate, solve, interpret, validate and so on), but the shape is always a loop:

  1. Define the question in a way that can be answered with numbers.
  2. Assume: decide what matters, what to ignore, and write it down.
  3. Build the mathematics: variables, equations, rules.
  4. Solve it, by hand or by computer.
  5. Check the result against units, limits, common sense and data.
  6. Report the answer, its limits and what to do next, then loop back.

The spirit is close to the four steps George Pólya taught for problem solving in 1945: understand the problem, make a plan, carry it out, look back. Modelling adds one crucial extra layer: the problem itself is messy, and you have to decide what it is.

Define

Sharp question Numbers, not feelings
Purpose Who will use the answer?
1 / 4

Why a cycle and not a line?

Suppose your coffee model predicts the drink reaches 60 °C after 9.5 minutes, but a stopwatch says 7. Something is off: perhaps heat also leaves through evaporation at the surface, which your first assumptions ignored. You return to step 2, add that effect, and rebuild. Each lap improves the model. The skill is knowing when it is good enough for the purpose, not when it is perfect.

A useful rule of thumb: do one very rough lap first. A crude model built in an hour often tells you which details matter, so the second lap spends effort where it counts.

What each lesson covers

Lesson You will learn
Defining the Problem and Choosing Assumptions Turning a vague wish into a question, sorting variables, parameters and constants, writing assumptions
Building the Model Balance laws, stocks and flows, constitutive rules, and how Newton's cooling law arises
Solving, Interpreting and Iterating Solving by formula and by code, reading the answer, sensitivity and the next lap
Chapter 2 Review A cheat sheet and a longer quiz

Try it

The fastest first lap is a Fermi estimate, named after the physicist Enrico Fermi, who was famous for quick order-of-magnitude answers. You split a big unknown into smaller guesses, give each a plausible range, and combine them. The answer is not exact but it tells you whether the true value is closer to ten, a thousand or a million.

Question: roughly how many liters of water does a city of one million people use for drinking and cooking in a day?

Python A Fermi estimate with ranges

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Now step through the cycle yourself. The stepper below walks one example through all six steps. Click "Next" and notice how the answer improves each time.

Walk the cycle: a stepper

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. Which is the correct order of the six steps?

    You pose the question, choose assumptions, build the mathematics, solve it, check it and report, then often loop.
  2. 2. Why is the modelling process drawn as a cycle?

    A failed check usually points back to an assumption or a variable, and the improved model gets a new lap.
  3. 3. What is a Fermi estimate?

    Fermi estimates give order-of-magnitude answers quickly, which helps you decide where to spend effort.
  4. 4. A good first lap of modelling should usually be:

    A crude model shows which parts of the problem the answer is sensitive to.

Summary

  • The modelling cycle has six steps: define, assume, build, solve, check, report.
  • It loops because checking reveals what to improve.
  • Do a rough lap first, then spend effort where it matters.
  • Fermi estimates are a fast way to start.
  • This chapter follows one example, a cooling cup of coffee, through every step.

Next: Defining the Problem and Choosing Assumptions