Krein–Milman theorem

In the mathematical theory of functional analysis, the Krein–Milman theorem is a proposition about compact convex sets in locally convex topological vector spaces (TVSs).

Given a convex shape (light blue) and its set of extreme points (red), the convex hull of is

Krein–Milman theorem[1] — A compact convex subset of a Hausdorff locally convex topological vector space is equal to the closed convex hull of its extreme points.

This theorem generalizes to infinite-dimensional spaces and to arbitrary compact convex sets the following basic observation: a convex (i.e. "filled") triangle, including its perimeter and the area "inside of it", is equal to the convex hull of its three vertices, where these vertices are exactly the extreme points of this shape. This observation also holds for any other convex polygon in the plane

Statement and definitions

edit

Preliminaries and definitions

edit
 
A convex set in light blue, and its extreme points in red.

Throughout,   will be a real or complex vector space.

For any elements   and   in a vector space, the set   is called the closed line segment or closed interval between   and   The open line segment or open interval between   and   is   when   while it is   when  [2] it satisfies   and   The points   and   are called the endpoints of these interval. An interval is said to be non-degenerate or proper if its endpoints are distinct.

The intervals   and   always contain their endpoints while   and   never contain either of their endpoints. If   and   are points in the real line   then the above definition of   is the same as its usual definition as a closed interval.

For any   the point   is said to (strictly) lie between   and   if   belongs to the open line segment  [2]

If   is a subset of   and   then   is called an extreme point of   if it does not lie between any two distinct points of   That is, if there does not exist   and   such that   and   In this article, the set of all extreme points of   will be denoted by  [2]

For example, the vertices of any convex polygon in the plane   are the extreme points of that polygon. The extreme points of the closed unit disk in   is the unit circle. Every open interval and degenerate closed interval in   has no extreme points while the extreme points of a non-degenerate closed interval   are   and  

A set   is called convex if for any two points     contains the line segment   The smallest convex set containing   is called the convex hull of   and it is denoted by   The closed convex hull of a set   denoted by   is the smallest closed and convex set containing   It is also equal to the intersection of all closed convex subsets that contain   and to the closure of the convex hull of  ; that is,   where the right hand side denotes the closure of   while the left hand side is notation. For example, the convex hull of any set of three distinct points forms either a closed line segment (if they are collinear) or else a solid (that is, "filled") triangle, including its perimeter. And in the plane   the unit circle is not convex but the closed unit disk is convex and furthermore, this disk is equal to the convex hull of the circle.

The separable Hilbert space Lp space   of square-summable sequences with the usual norm   has a compact subset   whose convex hull   is not closed and thus also not compact.[3] However, like in all complete Hausdorff locally convex spaces, the closed convex hull   of this compact subset will be compact.[4] But if a Hausdorff locally convex space is not complete then it is in general not guaranteed that   will be compact whenever   is; an example can even be found in a (non-complete) pre-Hilbert vector subspace of   Every compact subset is totally bounded (also called "precompact") and the closed convex hull of a totally bounded subset of a Hausdorff locally convex space is guaranteed to be totally bounded.[5]

Statement

edit

Krein–Milman theorem[6] — If   is a compact subset of a Hausdorff locally convex topological vector space then the set of extreme points of   has the same closed convex hull as  

In the case where the compact set   is also convex, the above theorem has as a corollary the first part of the next theorem,[6] which is also often called the Krein–Milman theorem.

Krein–Milman theorem[2] — Suppose   is a Hausdorff locally convex topological vector space (for example, a normed space) and   is a compact and convex subset of   Then   is equal to the closed convex hull of its extreme points:  

Moreover, if   then   is equal to the closed convex hull of   if and only if   where   is closure of  

The convex hull of the extreme points of   forms a convex subset of   so the main burden of the proof is to show that there are enough extreme points so that their convex hull covers all of   For this reason, the following corollary to the above theorem is also often called the Krein–Milman theorem.

(KM) Krein–Milman theorem (Existence)[2] — Every non-empty compact convex subset of a Hausdorff locally convex topological vector space has an extreme point; that is, the set of its extreme points is not empty.

To visualized this theorem and its conclusion, consider the particular case where   is a convex polygon. In this case, the corners of the polygon (which are its extreme points) are all that is needed to recover the polygon shape. The statement of the theorem is false if the polygon is not convex, as then there are many ways of drawing a polygon having given points as corners.

The requirement that the convex set   be compact can be weakened to give the following strengthened generalization version of the theorem.[7]

(SKM) Strong Krein–Milman theorem (Existence)[8] — Suppose   is a Hausdorff locally convex topological vector space and   is a non-empty convex subset of   with the property that whenever   is a cover of   by convex closed subsets of   such that   has the finite intersection property, then   is not empty. Then   is not empty.

The property above is sometimes called quasicompactness or convex compactness. Compactness implies convex compactness because a topological space is compact if and only if every family of closed subsets having the finite intersection property (FIP) has non-empty intersection (that is, its kernel is not empty). The definition of convex compactness is similar to this characterization of compact spaces in terms of the FIP, except that it only involves those closed subsets that are also convex (rather than all closed subsets).

More general settings

edit

The assumption of local convexity for the ambient space is necessary, because James Roberts (1977) constructed a counter-example for the non-locally convex space   where  [9]

Linearity is also needed, because the statement fails for weakly compact convex sets in CAT(0) spaces, as proved by Nicolas Monod (2016).[10] However, Theo Buehler (2006) proved that the Krein–Milman theorem does hold for metrically compact CAT(0) spaces.[11]

edit

Under the previous assumptions on   if   is a subset of   and the closed convex hull of   is all of   then every extreme point of   belongs to the closure of   This result is known as Milman's (partial) converse to the Krein–Milman theorem.[12]

The Choquet–Bishop–de Leeuw theorem states that every point in   is the barycenter of a probability measure supported on the set of extreme points of  

Relation to the axiom of choice

edit

Under the Zermelo–Fraenkel set theory (ZF) axiomatic framework, the axiom of choice (AC) suffices to prove all versions of the Krein–Milman theorem given above, including statement KM and its generalization SKM. The axiom of choice also implies, but is not equivalent to, the Boolean prime ideal theorem (BPI), which is equivalent to the Banach–Alaoglu theorem. Conversely, the Krein–Milman theorem KM together with the Boolean prime ideal theorem (BPI) imply the axiom of choice.[13] In summary, AC holds if and only if both KM and BPI hold.[8] It follows that under ZF, the axiom of choice is equivalent to the following statement:

The closed unit ball of the continuous dual space of any real normed space has an extreme point.[8]

Furthermore, SKM together with the Hahn–Banach theorem for real vector spaces (HB) are also equivalent to the axiom of choice.[8] It is known that BPI implies HB, but that it is not equivalent to it (said differently, BPI is strictly stronger than HB).

History

edit

The original statement proved by Mark Krein and David Milman (1940) was somewhat less general than the form stated here.[14]

Earlier, Hermann Minkowski (1911) proved that if   is 3-dimensional then   equals the convex hull of the set of its extreme points.[15] This assertion was expanded to the case of any finite dimension by Ernst Steinitz (1916).[16] The Krein–Milman theorem generalizes this to arbitrary locally convex  ; however, to generalize from finite to infinite dimensional spaces, it is necessary to use the closure.

See also

edit

Citations

edit
  1. ^ Rudin 1991, p. 75 Theorem 3.23.
  2. ^ a b c d e Narici & Beckenstein 2011, pp. 275–339.
  3. ^ Aliprantis & Border 2006, p. 185.
  4. ^ Trèves 2006, p. 145.
  5. ^ Trèves 2006, p. 67.
  6. ^ a b Grothendieck 1973, pp. 187–188.
  7. ^ Pincus 1974, pp. 204–205.
  8. ^ a b c d Bell, J. L.; Jellett, F. (1971). "On the Relationship Between the Boolean Prime Ideal Theorem and Two Principles in Functional Analysis" (PDF). Bull. Acad. Polon. Sci. sciences math., astr. et phys. 19 (3): 191–194. Retrieved 23 Dec 2021.
  9. ^ Roberts, J. (1977), "A compact convex set with no extreme points", Studia Mathematica, 60 (3): 255–266, doi:10.4064/sm-60-3-255-266
  10. ^ Monod, Nicolas (2016), "Extreme points in non-positive curvature", Studia Mathematica, 234: 265–270, arXiv:1602.06752
  11. ^ Buehler, Theo (2006), The Krein–Mil'man theorem for metric spaces with a convex bicombing, arXiv:math/0604187, Bibcode:2006math......4187B
  12. ^ Milman, D. (1947), Характеристика экстремальных точек регулярно-выпуклого множества [Characteristics of extremal points of regularly convex sets], Doklady Akademii Nauk SSSR (in Russian), 57: 119–122
  13. ^ Bell, J.; Fremlin, David (1972). "A geometric form of the axiom of choice" (PDF). Fundamenta Mathematicae. 77 (2): 167–170. doi:10.4064/fm-77-2-167-170. Retrieved 11 June 2018. Theorem 1.2. BPI [the Boolean Prime Ideal Theorem] & KM [Krein-Milman]   (*) [the unit ball of the dual of a normed vector space has an extreme point].... Theorem 2.1. (*)   AC [the Axiom of Choice].
  14. ^ Krein, Mark; Milman, David (1940), "On extreme points of regular convex sets", Studia Mathematica, 9: 133–138, doi:10.4064/sm-9-1-133-138
  15. ^ Minkowski, Hermann (1911), Gesammelte Abhandlungen, vol. 2, Leipzig: Teubner, pp. 157–161
  16. ^ Steinitz, Ernst (1916), "Bedingt konvergente Reihen und konvexe Systeme VI, VII", J. Reine Angew. Math., 146: 1–52, doi:10.1515/crll.1916.146.1, S2CID 122897233; (see p. 16)

Bibliography

edit

This article incorporates material from Krein–Milman theorem on PlanetMath, which is licensed under the Creative Commons Attribution/Share-Alike License.