Tarski-Knaster theorem


The aim of this article is to prove

Theorem 1 (LATTICE-THEORETICAL FIXPOINT THEOREM).

Let

  1. (i)

    𝔄=(A,)  be a complete latticeMathworldPlanetmath,

  2. (ii)

    f be an increasing function on A to A,

  3. (iii)

    P the set of all fixpoints of f.

Then the set P is not empty and the system (P,) is a complete lattice; in particular we have

P=Ex[f(x)x]P

and

P=Ex[f(x)x]P.

This article follows, with clarification, Tarski’s original 1955 article [6], which may not be available to some readers.

1 Definitions

The symbol denotes the empty setMathworldPlanetmath. The symbol is shorthand for ”for all”.

Definition 1.

A poset (partially ordered setMathworldPlanetmath) A=(A,) consists of a nonempty set A and a relationMathworldPlanetmath

:A×A{𝐹𝑎𝑙𝑠𝑒,𝑇𝑟𝑢𝑒}

that is reflexiveMathworldPlanetmathPlanetmath (aa aA), antisymmetric (a,bA, if ab and ba, then a=b). and transitiveMathworldPlanetmathPlanetmathPlanetmath  (a,b,cA, if ab and bc, then ac).

Definition 2.

If A=(A,) is a poset and B is a nonempty subset of A, then sup(B) is an element of A, if it exists, with these properties:

  1. 1.

    bB,bsup(B), (in shorthand, Bsup(B)).

  2. 2.

    If aA and Ba, then sup(B)a.

Similarly inf(B) is an element of A, if it exists, with these properties:

  1. (3)

    inf(B)B,

  2. (4)

    If aA and aB, then ainf(B).

Theorem 2.

If sup(B) exists, then it is unique; if inf(B) exists, it is unique.

Definition 3.

Let A=(A,) be a poset. Then A is a latticeMathworldPlanetmath if a,bA, the set {a,b} has a sup and an inf. We write

ab=sup({a,b})𝑎𝑛𝑑ab=inf({a,b}).

As an exercise, prove the following:

Theorem 3.

Let A=(A,) be a lattice. Then

  1. 1.

    ab=ba,ab=ba

  2. 2.

    (ab)c=a(bc),(ab)c=a(bc)

  3. 3.

    a(bc)(ab)(ac).  (ab)(ac)a(bc).

Definition 4.

A poset A=(A,) is a complete lattice if for each nonempty BA, both sup(B) and inf(B) exist.

In particular, this is the case when B={a,b}, so a complete lattice is a lattice.

2 The Tarski-Knaster Theorem

Statement of the theoremMathworldPlanetmath, as it is in [6]:

Theorem 4 (LATTICE-THEORETICAL FIXPOINT THEOREM).

Let

  1. (i)

    𝔄=(A,)  be a complete lattice,

  2. (ii)

    f be an increasing function on A to A,

  3. (iii)

    P the set of all fixpoints of f.

Then the set P is not empty and the system (P,) is a complete lattice; in particular we have

P=Ex[f(x)x]P

and

P=Ex[f(x)x]P.

The symbols P and P are what I am calling sup(P) and inf(C). Now let me restate the theorem:

Theorem 5.

Let A=(A,) be a complete lattice. Let f be an increasing function on A to A, that is, whenever ab, then f(a)f(b). Let P be the set of all fixed pointsPlanetmathPlanetmath of f:

P={aA|f(a)=a}.

Then

  1. 1.

    P.

  2. 2.

    The system 𝔓=(P, ) is a complete lattice.

  3. 3.

    Set B={bA|bf(b)}. Then B and sup(B)=sup(P)P.

  4. 4.

    Set C={cA|f(c)c}. Then Cand inf(C)=inf(P)P.

Lemma 1.

Let A=(A,) be a complete lattice and A0A. Let

D={dA|dA0},E={eA|A0e}.

(dAo means da0 a0A0.) Then both B and C are completePlanetmathPlanetmathPlanetmathPlanetmath lattices, with  inherited from that of A.

Proof (of Lemma).

We’ll do the case of D; that for E is similar (dual). First, D because inf(A)D. Suppose GD. Obviously

inf(G)gA0

for any gG, so inf(G)D. As gA0 for all gG, we have sup(G)A0 by (2) in the definition of sup. So sup(G)D. As G was any nonempty subset of D, these statements say that D is a complete lattice. ∎

Proof (of Theorem).

B because inf(A)B. Hence b1=sup(B) exists. If bB, then bb1, which implies

f(b)f(b1).

because f is increasing., so

bf(b)f(b1).

This is true bB. Hence

b1=sup(B)f(b1).

Therefore b1B. Because f is increasing, this implies

f(b1)f(f(b1)),

from which f(b1)B. Hence f(b1)sup(B)=b1. From

b1f(b1)b1

we have f(b1)=b1, that is, b1P. This settles (1): P, and incidentally proves sup(B)P.

From this, sup(B)sup(P). But obviously PB, so sup(P)sup(B). Therefore sup(P)=sup(B), completing the proof of (3). Statement (4) is proved similarly (or dually).

Next we prove (2). Let P0P. Set

Q={qA|qP0}.

By the Lemma, Q is a complete lattice. If qQ, then p0P0, we have qp0, hence

f(q)f(p0)=p0,

that is, f(q)Q. In other words,

f:QQ.

By what we have already proved for A, applied to Q, we have sup(Q)Q, and sup(Q) is a fixed point of f. That is, sup(Q)P. But obviously sup(Q)=inf(P0), so inf(P0)P Similarly sup(P0)P, completing the proof of (2): P is a complete lattice. ∎

This theorem was proved by A. Tarski [6]. A special case of this theorem (for lattices of sets) appeared in a paper of B. Knaster [3]. Kind of converseMathworldPlanetmath of this theorem was proved by Anne C. Davis [1]: If every order-preserving function f:LL has a fixed point, then L is a complete lattice.

References

  • 1 Anne C. Davis, A characterizationMathworldPlanetmath of complete lattices., Pac. J. Math. 5 (1955), 311–319.
  • 2 Thomas Forster, Logic, inductionMathworldPlanetmath and sets, Cambridge University Press, Cambridge, 2003.
  • 3 B. Knaster, Un théorème sur les fonctions d’ensembles., Annales Soc. Polonaise 6 (1928), 133–134.
  • 4 M. Kolibiar, A. Legéň, T. Šalát, and Š. Znám, Algebra a príbuzné disciplíny, Alfa, Bratislava, 1992 (Slovak).
  • 5 J. B. Nation: http://bigcheese.math.sc.edu/ mcnulty/alglatvar/Notes on lattice theory
  • 6 Alfred Tarski, A lattice-theoretical fixpoint theorem and its applications., Pac. J. Math. 5 (1955), 285–309.
  • 7 Wikipedia’s entry on http://en.wikipedia.org/wiki/Knaster-Tarski_theoremKnaster-Tarski theorem
Title Tarski-Knaster theorem
Canonical name TarskiKnasterTheorem
Date of creation 2013-03-22 15:30:19
Last modified on 2013-03-22 15:30:19
Owner kompik (10588)
Last modified by kompik (10588)
Numerical id 18
Author kompik (10588)
Entry type Theorem
Classification msc 06B99
Synonym Tarski theoremMathworldPlanetmath
Synonym Knaster-Tarski theorem
Related topic lattice
Related topic completelattice
Related topic fixedpoint
Related topic ProofOfSchroederBernsteinTheoremUsingTarskiKnasterTheorem
Related topic Lattice
Related topic CompleteLattice