TheoremComplete

Brouwer Fixed-Point Theorem (Dimension 2)

The Brouwer fixed-point theorem states that every continuous map from the closed disk to itself has a fixed point. The two-dimensional case can be proved using the fundamental group, making it one of the first applications of algebraic topology to a concrete analytical question.


Statement

Theorem9.11Brouwer Fixed-Point Theorem (2D)

Every continuous map f:D2D2f: D^2 \to D^2 has a fixed point: there exists xD2x \in D^2 with f(x)=xf(x) = x.

Here D2={(x,y)R2:x2+y21}D^2 = \{(x, y) \in \mathbb{R}^2 : x^2 + y^2 \leq 1\} is the closed unit disk.


Proof via the Fundamental Group

Proof

The proof proceeds by contradiction, using two key lemmas.

Lemma 1 (No Retraction): There is no retraction r:D2S1r: D^2 \to S^1, i.e., no continuous map rr with r(x)=xr(x) = x for all xS1x \in S^1.

Proof of Lemma 1: Suppose r:D2S1r: D^2 \to S^1 is a retraction. Let ι:S1D2\iota: S^1 \hookrightarrow D^2 be the inclusion. Then rι=idS1r \circ \iota = \operatorname{id}_{S^1}, so (rι)=rι=idπ1(S1)(r \circ \iota)_* = r_* \circ \iota_* = \operatorname{id}_{\pi_1(S^1)}.

Now ι:π1(S1)π1(D2)\iota_*: \pi_1(S^1) \to \pi_1(D^2) must be injective (if rιr_* \circ \iota_* is the identity, ι\iota_* is injective). But π1(S1)Z\pi_1(S^1) \cong \mathbb{Z} and π1(D2)=0\pi_1(D^2) = 0 (since D2D^2 is contractible), so no injective homomorphism Z0\mathbb{Z} \to 0 exists. Contradiction.

Lemma 2 (Fixed-point-free maps yield retractions): If f:D2D2f: D^2 \to D^2 has no fixed point, then there exists a retraction r:D2S1r: D^2 \to S^1.

Proof of Lemma 2: For each xD2x \in D^2, since f(x)xf(x) \neq x, there is a unique ray starting at f(x)f(x) and passing through xx. This ray hits S1S^1 at a unique point; call it r(x)r(x).

Explicitly: r(x)r(x) is the point on S1S^1 obtained by starting at f(x)f(x), moving in the direction xf(x)x - f(x), and hitting the boundary. Analytically: r(x)=x+t(x)(xf(x))r(x) = x + t(x)(x - f(x)) where t(x)0t(x) \geq 0 is chosen so that r(x)=1\|r(x)\| = 1. This is the larger root of a quadratic in tt, which depends continuously on xx (since f(x)xf(x) \neq x ensures the ray is well-defined).

For xS1x \in S^1: since x=1\|x\| = 1 and f(x)D2f(x) \in D^2, the ray from f(x)f(x) through xx exits D2D^2 at xx itself (with t=0t = 0 giving r(x)=xr(x) = x). So rS1=idS1r|_{S^1} = \operatorname{id}_{S^1}.

Combining: If f:D2D2f: D^2 \to D^2 has no fixed point, Lemma 2 gives a retraction r:D2S1r: D^2 \to S^1, contradicting Lemma 1. Therefore ff must have a fixed point.


Applications

ExampleApplications of Brouwer's Theorem
  1. Existence of equilibria: In game theory, Brouwer's theorem (and its generalization, Kakutani's theorem) is used to prove the existence of Nash equilibria.

  2. Nonlinear equations: If f:D2D2f: D^2 \to D^2 is continuous, the equation f(x)=xf(x) = x always has a solution. This is used in PDE theory (via Schauder's fixed-point theorem, which generalizes Brouwer to infinite dimensions).

  3. The Ham Sandwich Theorem and other results in combinatorial geometry ultimately rely on fixed-point or degree-theoretic arguments.

ExampleNon-Constructive Nature

The Brouwer theorem guarantees existence of a fixed point but provides no algorithm for finding it. Consider f:D2D2f: D^2 \to D^2 defined by a complicated continuous formula. We know ff has a fixed point, but finding it may require iterative methods (not guaranteed to converge in general).

This non-constructive character led Brouwer himself to become an advocate of intuitionism, rejecting non-constructive existence proofs!


Generalizations

Theorem9.12Brouwer Fixed-Point Theorem (General)

Every continuous map f:DnDnf: D^n \to D^n (the closed nn-ball in Rn\mathbb{R}^n) has a fixed point, for every n1n \geq 1.

RemarkProofs in Higher Dimensions

The proof for n=2n = 2 uses π1\pi_1. For general nn, one needs stronger tools:

  • Homology theory: The proof generalizes by replacing π1\pi_1 with Hn1H_{n-1} (the (n1)(n-1)-th homology group). The no-retraction lemma becomes: Hn1(Sn1)ZH_{n-1}(S^{n-1}) \cong \mathbb{Z} does not inject into Hn1(Dn)=0H_{n-1}(D^n) = 0.
  • Degree theory: The degree of idSn1\operatorname{id}_{S^{n-1}} is 101 \neq 0, but if a retraction existed, the degree would factor through Hn1(Dn)=0H_{n-1}(D^n) = 0.
  • Smooth methods: In the smooth category, Sard's theorem and differential forms provide alternative proofs.
Theorem9.13Borsuk--Ulam Theorem (2D)

For every continuous map f:S2R2f: S^2 \to \mathbb{R}^2, there exists a point xS2x \in S^2 with f(x)=f(x)f(x) = f(-x). That is, some pair of antipodal points maps to the same value.

RemarkConnection to Brouwer

The Borsuk--Ulam theorem implies the Brouwer fixed-point theorem (in dimension 2) and is itself stronger. It has a memorable consequence: at any moment, there are two antipodal points on the Earth's surface with the same temperature and the same atmospheric pressure.