Meaning and direct answer
Fixed-point iteration solves x=g(x) through repeated substitution. A contraction on a complete invariant set guarantees a unique fixed point and convergence from every permitted starting value.[1]
Fixed-point iteration solves x=g(x) through repeated substitution. A contraction on a complete invariant set guarantees a unique fixed point and convergence from every permitted starting value.
Assumptions and limits
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.[1]
Why the relation holds
Step sizes shrink at least geometrically. Their remaining sum is bounded by a geometric series. Completeness provides a limit in the set, and continuity makes it a fixed point. Two fixed points would satisfy d≤qd and therefore coincide.[1]
The model visualized
Iteration (x+2)/3 starting at x0=0
Free with a Quanta account. Your download stays ready after registration.
Two fully worked examples
Affine contraction
- Given
- g(x)=(x+2)/3 on I=[0,2], x0=0.
- Find
- Answer the stated question using valid steps and verify the result.
Step sizes shrink at least geometrically. Their remaining sum is bounded by a geometric series. Completeness provides a limit in the set, and continuity makes it a fixed point. Two fixed points would satisfy d≤qd and therefore coincide.
Substitute the stated data or evaluate the defining relation.
Result: After two steps the true error is 1/9.
Check: q/(1−q)·|8/9−2/3|=1/9, an exact bound in this case.
Same solution, poor iteration
- Given
- x=1 is a fixed point of g(x)=2x−1. Start x0=1.1.
- Find
- Answer the stated question using valid steps and verify the result.
Step sizes shrink at least geometrically. Their remaining sum is bounded by a geometric series. Completeness provides a limit in the set, and continuity makes it a fixed point. Two fixed points would satisfy d≤qd and therefore coincide.
Substitute the stated data or evaluate the defining relation.
Result: Error grows; the mapping is not a contraction.
Check: The equation can be valid even though this iteration diverges.
Recall the method later with the formula’s flashcards.
10 flashcards · English materials
Already have an account? Sign in for flashcardsAvailable from Essential. A ready-made formula set, no AI credits required. Article exercises are not all included.
A priori bound and number of steps
Under the hypotheses of the Banach fixed-point theorem (closed set, self-map, 0 < q < 1); ε > 0 is the required accuracy and x1 ≠ x0.[2][3]
The step sizes satisfy . Summing all later steps as a geometric series gives the a priori bound. It needs only the first step and therefore tells you before computing how many steps are enough. Because ln q is negative, the inequality sign flips when solving for n. The a posteriori bound with the last step is never worse, but it is available only during the computation.[2][3]
Example: maps [0.4, 0.7] into [0.4966, 0.6703], and . With x0 = 0.5 we get and |x1 − x0| ≈ 0.1065. For ε = 10−4 the bound requires n ≥ 20.2, so 21 steps. In fact x12 ≈ 0.56707 is already closer than 10−4 to the fixed point 0.56714, and the a posteriori bound drops below 10−4 after 15 steps. The a priori bound is therefore safe but cautious, because q is the maximum of |g′| over the whole interval; near the fixed point only acts.[2][3]
The sign of g′ shapes the sequence. Here , so the values alternate between the two sides of the fixed point: x0 = 0.5 lies to the left, x1 ≈ 0.6065 to the right, x2 ≈ 0.5452 to the left again. If instead 0 < g′ < 1, the sequence approaches monotonically from one side.[3]
Five exercises with solutions
Calculate · Foundation
Find the fixed point of g(x)=0.5x+1.
Show hint
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.
Learning goal: Find the fixed point of g(x)=0.5x+1.
Show solution and calculation
x=0.5x+1 gives x=2.
Rearrange · Foundation
q=0.2 and last step 0.004: find the a-posteriori error bound.
Show hint
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.
Learning goal: q=0.2 and last step 0.004: find the a-posteriori error bound.
Show solution and calculation
q/(1−q)·0.004=0.001.
Interpret · Application
Why must the set be invariant?
Show hint
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.
Learning goal: Why must the set be invariant?
Show solution and calculation
Otherwise an iterate can leave the region where the bound and validity were proved.
Error analysis · Application
Can q=1 be used in the contraction theorem?
Show hint
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.
Learning goal: Can q=1 be used in the contraction theorem?
Show solution and calculation
No.
The geometric error tail is uncontrolled; g(x)=x has infinitely many fixed points.
Transfer · Application
An iteration alternates between two values. Does this prove the equation has no solution?
Show hint
For the interval version, I is a nonempty closed real interval, g maps I into I, and |g(x)−g(y)|≤q|x−y| for one fixed 0≤q<1. A derivative bound |g′|≤q suffices. Merely checking |g′(x0)|<1 at one point does not.
Learning goal: An iteration alternates between two values. Does this prove the equation has no solution?
Show solution and calculation
No.
The iteration can be unsuitable, as with g(x)=1−x despite its fixed point 1/2.
Check your understanding with the formula’s reference test.
8 test questions · English materials
Already have an account? Sign in for the testAvailable from Essential. A ready-made formula set, no AI credits required. Article exercises are not all included.
Related next learning steps
Related methods: Newton’s method for finding roots, Convergence Tests for Infinite Series.
Sources for this explanation
Quanta · Updated November 6, 2026
- Jiří Lebl, Basic Analysis I, 7.6 Contraction Mapping Principle (MIT OpenCourseWare edition)
- RWTH Aachen, IGPM (German): Numerical mathematics for mechanical engineers, nonlinear systems I (fixed-point iteration, Banach fixed-point theorem, a priori and a posteriori error estimates), 2018
- David Bindel, Cornell CS 4220 Numerical Analysis: Fixed point iteration and Newton (contraction mappings, error bound, asymptotic convergence), 2026
Editorial responsibility: Quanta Study editorial team
Review scope
This explanation, 2 worked examples and 5 exercises are available in German and English. Structure, source references and mathematical notation undergo automated checks. Calculations, model assumptions and both language versions receive a separate AI review.
An independent, named human subject review has not yet been documented. Our exercise numbers are worked examples, rather than measurements published by the cited sources.
Editorial responsibility, sources and methodsChange history
- Authored a German-English topic with assumptions, reasoning, original examples and five solved exercises.
Cite this reference page
Quanta. Fixed-point iteration and contraction. Updated November 6, 2026. https://quanta-study.de/en/formula-sheet/fixpunktiteration-banach
When citing a specific scientific statement, also cite the specialist source linked to that statement.