Iterative & Numerical Methods — AHL 2.9 Iterative methods and numerical solutions using technology (HL extension) - IB Mathematics Applications and Interpretation

SECTION 1

Historical Context & Motivation

Mathematics has always aimed to solve equations exactly, but many equations resist neat, closed-form solutions. As early as ancient Babylon, mathematicians discovered that repeatedly refining a guess could lead to remarkably accurate answers. This approach — called iteration — means applying a process again and again, each time getting closer to the true answer. For centuries, these methods were tedious to carry out by hand, but the rise of computers and graphing calculators transformed them into powerful, practical tools.

~2000 BCE • Babylonian Square Roots

Babylonian scribes used an iterative averaging method to approximate square roots on clay tablets — the same idea behind the algorithm you may know as Heron's method.

1669 • Newton's Method

Isaac Newton developed an iterative technique using tangent lines to find roots of equations. Later refined by Joseph Raphson, it became the celebrated Newton–Raphson method.

1795 • Fixed-Point Iteration Formalized

Mathematicians formalized the idea of rewriting an equation as x = g(x) and iterating. Conditions for convergence were studied, laying the groundwork for modern fixed-point iteration.

1950s • The Computer Revolution

Electronic computers made thousands of iterations possible in seconds. Numerical methods moved from theoretical curiosity to essential engineering and scientific practice.

Today • GDC & CAS in the Classroom

Graphing display calculators (GDCs) and computer algebra systems let IB students apply iterative methods to solve equations that resist algebraic techniques.

The central question that drives this topic is straightforward: When an equation has no exact algebraic solution, how can we still find accurate numerical answers? Iterative methods provide the answer by turning guessing into a systematic, convergent process.

SECTION 2

Core Principles & Definitions

Before diving into calculations, it helps to understand the foundational ideas that make iterative methods work. These principles apply whether you are solving equations by hand, on a GDC, or with software.

Iteration

Repeating a calculation using the output of one step as the input for the next. Each cycle is called an iterate, and the sequence of iterates (x₀, x₁, x₂, …) ideally approaches the solution.

Fixed Point

A value x* such that g(x*) = x*. When the iteration formula x_n+1 = g(x_n) converges, it settles at this fixed point.

Convergence vs. Divergence

A sequence converges when successive iterates get closer to the solution. It diverges when they move away — meaning the rearrangement or starting value must be changed.

Graphical Interpretation

A root of f(x) = 0 is where the graph crosses the x-axis. This graphical view helps you pick a good initial estimate (x₀) before iterating.

Technology as a Tool

Your GDC's equation solver, table function, or spreadsheet can execute dozens of iterations instantly, making numerical solutions practical and reliable in exam settings.

✦

KEY TAKEAWAY

Think of iteration like adjusting the temperature dial on a shower. You start with a guess, check the result, and adjust — each time getting closer to the perfect setting. In mathematics, the 'guess-and-adjust' cycle is automated by a formula, and the 'perfect setting' is the fixed point of that formula.

SECTION 3

Visual Explanation — The Cobweb Diagram

The best way to understand how iteration converges to a fixed point is through a cobweb diagram (also called a staircase diagram). This diagram plots both y = g(x) and y = x on the same axes. Starting from your initial guess x₀, you trace a vertical line up to the curve y = g(x), then a horizontal line to the line y = x, then repeat. The path spirals or staircases toward the fixed point where the curve and line intersect.

Starting from x₀, the pink cobweb path traces vertical steps to the curve y = g(x) and horizontal steps to the line y = x. The path spirals inward toward the fixed point (yellow dot), showing convergence.

In the diagram above, notice how each successive iterate (x₁, x₂, …) is closer to the fixed point than the one before. The key geometric insight is that convergence occurs when the curve y = g(x) is less steep than y = x near the fixed point. More precisely, the iteration converges when |g′(x*)| < 1 — that is, the slope of g at the fixed point has an absolute value less than one. If the curve is steeper than y = x at the crossing, the cobweb spirals outward and the iteration diverges.

SECTION 4

Mathematical Framework

The mathematics of iterative methods centers on rearranging an equation f(x) = 0 into the form x = g(x), then generating a sequence of approximations. Below are the core formulas you need for AHL 2.9.

FIXED-POINT ITERATION FORMULA

x_{n+1} = g(x_n), n = 0, 1, 2, …

x₀ is the initial estimate (your starting guess). Each x_n+1 is computed from the previous x_n by applying the function g.

CONVERGENCE CONDITION

|g′(x*)| < 1

The iteration converges to the fixed point x* provided the absolute value of the derivative g′ at x* is strictly less than 1. If |g′(x*)| ≥ 1, the iteration may diverge and you need to try a different rearrangement.

REARRANGEMENT STRATEGY

f(x) = 0 → x = g(x)

Isolate x on one side to create the iteration function g(x). For example, x³ + x − 5 = 0 can be rearranged to x = $(5 − x)^{1/3}$ or x = 5 − x³, each giving a different g(x) with different convergence behavior.

NEWTON–RAPHSON METHOD (for reference)

x_{n+1} = x_n − f(x_n) / f′(x_n)

A powerful special case of iteration where the tangent line is used to find a better approximation. While not always required in AHL 2.9, your GDC's equation solver often uses this method internally.

💡

GDC Tip

On most graphing calculators, you can set up iteration using the Ans key. Enter your initial guess, press ENTER, then type the formula using Ans in place of x_n. Each press of ENTER generates the next iterate. Alternatively, use the TABLE or SEQUENCE mode for a full list of values.

SECTION 5

Convergence, Divergence & Choosing the Right Rearrangement

Not every rearrangement of an equation will produce a convergent iteration. The same equation f(x) = 0 can be rewritten as x = g(x) in multiple ways, and some rearrangements converge beautifully while others send your iterates flying off to infinity. Understanding why this happens — and how to choose the right form — is one of the most important skills in AHL 2.9.

Left panel: when the curve y = g(x) is shallower than y = x at the fixed point, the cobweb spirals inward (convergent). Right panel: when the curve is steeper than y = x, the cobweb spirals outward (divergent).

How to Choose a Good Rearrangement

Consider the equation x³ + x − 5 = 0. You could rearrange it as x = 5 − x³, giving g(x) = 5 − x³. Near the root (around x ≈ 1.5), the derivative g′(x) = −3x² ≈ −6.75, so |g′| ≈ 6.75 > 1 — this will diverge. Instead, rearranging to x = ∛(5 − x) gives g(x) = $(5 − x)^{1/3}$ , with g′(x) = −1/(3 $(5 − x)^{2/3}$ ) ≈ −0.13 near the root, so |g′| ≈ 0.13 < 1 — this will converge. When in doubt, graph both y = g(x) and y = x on your GDC and check which crossing has a shallower curve.

Rearrangement	g(x)	\|g′(x*)\| (approx.)	Converges?
x = 5 − x³	5 − x³	6.75	No
x = ∛(5 − x)	$(5 − x)^{1/3}$	0.13	Yes ✓
x = 5/(x² + 1)	5/(x² + 1)	0.65	Yes ✓

Comparing rearrangements of x³ + x − 5 = 0

SECTION 6

Worked Example

Let's work through a complete example using fixed-point iteration to solve an equation that cannot be solved algebraically.

Solving $e^{−x}$ = x using iteration

1
Step 1 — Identify the equation and rearrange

We want to solve e^−x = x. This equation mixes an exponential and a linear term, so there is no algebraic solution. The equation is already in the form x = g(x) with g(x) = e^−x. We check convergence: g′(x) = −e^−x, and near x ≈ 0.57, |g′| ≈ 0.57 < 1, so this rearrangement will converge.

g(x) = $e^{−x}$ , |g′(x*)| < 1 ✓
2
Step 2 — Choose an initial estimate

Graph y = e^−x and y = x on your GDC. They intersect near x ≈ 0.5 to 0.6. We choose x₀ = 0.5 as our starting value.

x₀ = 0.5
3
Step 3 — Iterate using your GDC

On your GDC, enter 0.5, press ENTER, then type $e^{−Ans}$ and press ENTER repeatedly. Alternatively, use a spreadsheet or the sequence mode. The iterates are: x₁ = e^−0.5 = 0.6065, x₂ = e^−0.6065 = 0.5452, x₃ = 0.5797, x₄ = 0.5601, x₅ = 0.5712, x₆ = 0.5649, x₇ = 0.5685, x₈ = 0.5664, x₉ = 0.5676, x₁₀ = 0.5669.

The iterates are settling around 0.567
4
Step 4 — Determine the solution to required accuracy

By x₁₅, the iterates agree to four decimal places: 0.5671. The solution to three significant figures is x ≈ 0.567. This value is known as the Omega constant in mathematics (approximately 0.5671).

x ≈ 0.567 (3 s.f.)
5
Step 5 — Verify using the GDC's equation solver

Use the solve function on your GDC: solve( $e^{−x}$ − x = 0, x). The calculator returns x = 0.56714329..., confirming our iterative answer. Always verify when possible!

Verified: x = 0.5671 (4 d.p.)

SECTION 7

Strengths, Limitations & Method Comparison

Iterative methods are incredibly useful, but they are not perfect. Understanding their strengths and limitations helps you decide when to use them and how to avoid common pitfalls.

Aspect	Strength	Limitation
Scope	Can solve equations that have no algebraic solution (e.g., $e^x$ = 3x)	Only gives a numerical approximation, not an exact answer
Speed	With a GDC, hundreds of iterations take seconds	By hand, each iteration requires careful arithmetic
Convergence	When \|g′(x)\| < 1, convergence is guaranteed near x	Wrong rearrangement → divergence; wrong x₀ → converging to wrong root
Accuracy	Can achieve any desired level of precision by doing more iterations	Rounding errors on calculators can accumulate over many steps
Multiple Roots	Can find different roots by choosing different starting values	May miss a root if your initial estimate is too far away

Strengths and limitations of iterative methods

✦

KEY TAKEAWAY

Iterative methods are like GPS navigation for equations: they don't tell you the road map formula — they guide you turn by turn to the destination. The catch is that you need a reasonable starting location and a set of directions (the rearrangement) that actually leads somewhere. A bad rearrangement is like faulty GPS instructions that send you in circles or off a cliff.

SECTION 8

Connections to Advanced Theory & Other IB Topics

The iterative methods you learn in AHL 2.9 connect to several other areas of mathematics and the wider IB curriculum. The table below maps out these connections.

AHL 2.9 Concept	Connection	Where It Appears
Fixed-point iteration	Recursive sequences and series	AHL 1.12 — Sequences defined recursively
Graphical root-finding	Intersections of function graphs	SL 2.5 — Solving equations graphically
Newton–Raphson method	Tangent lines and derivatives	AHL 5.9 — Differential calculus
Convergence conditions	Limits and behavior of sequences	AHL 1.13 — Limits of sequences
Numerical solutions with technology	Modeling real-world problems	All HL topics — Internal Assessment

Cross-topic connections within IB Mathematics AI HL

Looking ahead, iterative methods are foundational in university-level numerical analysis, a branch of mathematics essential for engineering, physics simulations, machine learning, and financial modeling. The Newton–Raphson method, for instance, underpins many of the optimization algorithms used to train artificial neural networks. Mastering the core idea of iteration now gives you a head start in these fields.

ℹ

IB Exam Advice

In the exam, you may be asked to: (1) set up an iteration formula from a given equation, (2) perform several iterations by hand or on your GDC, (3) state the solution to a given number of decimal places or significant figures, and (4) explain whether a particular rearrangement converges or diverges. Always show your iteration values clearly in a table.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

Explain in your own words what it means for an iteration x_n+1 = g(x_n) to converge. Why does the condition |g′(x*)| < 1 matter?

Show answer

Convergence means that as you compute more and more iterates, the values x₁, x₂, x₃, … get closer and closer to a single value x*, the fixed point. The condition |g′(x*)| < 1 ensures that near the fixed point, the function g 'contracts' distances — each iterate is pulled closer to x* rather than pushed away. If |g′(x*)| > 1, the function 'stretches' distances and iterates move further apart, causing divergence. Visually, the convergence condition means the curve y = g(x) crosses the line y = x at an angle that is less steep than 45°.

Question 2: BASIC CALCULATION

The equation x² − 3 = 0 is rearranged to x = (x + 3/x)/2. Starting with x₀ = 2, perform three iterations and give x₃ to five decimal places.

Show answer

Using g(x) = (x + 3/x)/2: x₁ = (2 + 3/2)/2 = (2 + 1.5)/2 = 1.75. x₂ = (1.75 + 3/1.75)/2 = (1.75 + 1.71429)/2 = 1.73214 (to 5 d.p.). x₃ = (1.73214 + 3/1.73214)/2 = (1.73214 + 1.73196)/2 = 1.73205. This converges rapidly to √3 ≈ 1.73205. This is actually Heron's method for square roots, a special case of fixed-point iteration that converges very quickly because the derivative of g at x* is zero.

Question 3: INTERMEDIATE

The equation x³ + 2x − 7 = 0 has a root near x = 1.4. (a) Show that the equation can be rearranged to x = (7 − x³)/2. (b) Explain why the rearrangement x = ∛(7 − 2x) is a better choice for iteration. (c) Use x₀ = 1.4 with the better rearrangement and perform four iterations, giving your answer to 4 decimal places.

Show answer

(a) Rearranging: x³ + 2x − 7 = 0 → 2x = 7 − x³ → x = (7 − x³)/2. (b) For g(x) = (7 − x³)/2, g′(x) = −3x²/2. At x ≈ 1.4, |g′| ≈ 3(1.96)/2 = 2.94 > 1, so this diverges. For g(x) =

(7 − 2x)^{1/3}

, g′(x) = −2/(3

(7 − 2x)^{2/3}

). At x ≈ 1.4, |g′| ≈ 2/(3 ×

(4.2)^{2/3}

) ≈ 2/(3 × 2.61) ≈ 0.26 < 1, so this converges. (c) x₀ = 1.4, x₁ =

(7 − 2.8)^{1/3}

4.2^{1/3}

= 1.6134, x₂ =

(7 − 3.2268)^{1/3}

3.7732^{1/3}

= 1.5567, x₃ =

(7 − 3.1134)^{1/3}

3.8866^{1/3}

= 1.5714, x₄ =

(7 − 3.1428)^{1/3}

3.8572^{1/3}

= 1.5676. The root is approximately x ≈ 1.5676 to 4 d.p.

Question 4: APPLIED

A population model predicts that the steady-state population P (in thousands) satisfies the equation P = 5 × ln(P + 2). Use your GDC with x₀ = 8 to find the steady-state population to 3 significant figures. How many iterations does it take for the answer to stabilize to 3 significant figures?

Show answer

Using g(P) = 5 × ln(P + 2) with P₀ = 8: P₁ = 5 × ln(10) = 5 × 2.3026 = 11.513. P₂ = 5 × ln(13.513) = 5 × 2.6036 = 13.018. P₃ = 5 × ln(15.018) = 5 × 2.7092 = 13.546. P₄ = 5 × ln(15.546) = 5 × 2.7437 = 13.719. P₅ = 5 × ln(15.719) = 5 × 2.7548 = 13.774. P₆ = 5 × ln(15.774) = 5 × 2.7583 = 13.791. P₇ = 13.798. P₈ = 13.800. The steady-state population is approximately 13,800 (13.8 thousand) to 3 significant figures. The answer stabilizes at 3 s.f. by around iteration 6–7. We can verify: g′(P) = 5/(P + 2); at P ≈ 13.8, |g′| ≈ 5/15.8 ≈ 0.32 < 1, confirming convergence.

Question 5: CRITICAL THINKING

The equation cos(x) = x has exactly one real root. (a) Explain why a root must exist between x = 0 and x = 1. (b) A student uses the rearrangement x = cos(x) with x₀ = 0 and another student uses x = arccos(x) with x₀ = 0.5. Both converge. Investigate which converges faster by computing five iterates of each. (c) Explain your findings in terms of |g′(x*)| for each method.

Show answer

(a) Let f(x) = cos(x) − x. At x = 0, f(0) = 1 − 0 = 1 > 0. At x = 1, f(1) = cos(1) − 1 ≈ 0.5403 − 1 = −0.4597 < 0. Since f is continuous and changes sign, by the Intermediate Value Theorem, a root exists between 0 and 1. (b) Method 1: g(x) = cos(x), x₀ = 0. x₁ = 1, x₂ = 0.5403, x₃ = 0.8576, x₄ = 0.6543, x₅ = 0.7935. Method 2: g(x) = arccos(x), x₀ = 0.5. x₁ = 1.0472, but arccos is only valid for |x| ≤ 1, and cos(x) = x has root ≈ 0.7391, so actually using g(x) = cos(x) is the standard approach. With Method 1, values oscillate and converge slowly. The root is x* ≈ 0.7391. (c) For g(x) = cos(x), g′(x) = −sin(x), so |g′(x*)| = sin(0.7391) ≈ 0.6736. Since this is less than 1 but not very small, convergence is moderate. A smaller value of |g′| would give faster convergence. This illustrates that convergence speed depends directly on how close |g′(x*)| is to zero — the closer to zero, the faster the method converges.

SUMMARY

Lesson Summary

Summary

In this lesson, you learned that iterative methods provide a systematic way to solve equations that resist algebraic techniques. The process begins by rearranging f(x) = 0 into the form x = g(x), choosing an initial estimate x₀ (informed by a graph), and applying the fixed-point iteration formula $x_{n+1}$ = g( $x_n$ ) repeatedly until the values stabilize to the required accuracy.

The convergence condition |g′(x*)| < 1 determines whether a rearrangement will work — visually captured by the cobweb diagram where convergent iterations spiral inward and divergent ones spiral outward. Your GDC is essential for performing iterations efficiently, graphing functions to locate roots, and verifying answers with built-in solvers. Remember: always show your iteration values clearly, state your answer to the required degree of accuracy, and check that your rearrangement satisfies the convergence condition.