This theorem has applications in physics when one describes the set of states of an algebra of observables, namely that any state can be written as a convex linear combination of so-called pure states.
According to Lawrence Narici and Edward Beckenstein, the Alaoglu theorem is a “very important result—maybe the most important fact about the weak-* topology—[that] echos throughout functional analysis.”[2] In 1912, Helly proved that the unit ball of the continuous dual space of is countably weak-* compact.[3] In 1932, Stefan Banach proved that the closed unit ball in the continuous dual space of any separablenormed space is sequentially weak-* compact (Banach only considered sequential compactness).[3] The proof for the general case was published in 1940 by the mathematician Leonidas Alaoglu. According to Pietsch [2007], there are at least twelve mathematicians who can lay claim to this theorem or an important predecessor to it.[2]
The Bourbaki–Alaoglu theorem is a generalization[4][5] of the original theorem by Bourbaki to dual topologies on locally convex spaces. This theorem is also called the Banach–Alaoglu theorem or the weak-* compactness theorem and it is commonly called simply the Alaoglu theorem.[2]
If is a vector space over the field then will denote the algebraic dual space of and these two spaces are henceforth associated with the bilinearevaluation map defined by
where the triple forms a dual system called the canonical dual system.
If is a topological vector space (TVS) then its continuous dual space will be denoted by where always holds. Denote the weak-* topology on by and denote the weak-* topology on by The weak-* topology is also called the topology of pointwise convergence because given a map and a net of maps the net converges to in this topology if and only if for every point in the domain, the net of values converges to the value
of any neighborhood of origin in is compact in the weak-* topology[note 1] on Moreover, is equal to the polar of with respect to the canonical system and it is also a compact subset of
To start the proof, some definitions and readily verified results are recalled. When is endowed with the weak-* topology then this Hausdorfflocally convex topological vector space is denoted by The space is always a complete TVS; however, may fail to be a complete space, which is the reason why this proof involves the space Specifically, this proof will use the fact that a subset of a complete Hausdorff space is compact if (and only if) it is closed and totally bounded. Importantly, the subspace topology that inherits from is equal to This can be readily verified by showing that given any a net in converges to in one of these topologies if and only if it also converges to in the other topology (the conclusion follows because two topologies are equal if and only if they have the exact same convergent nets).
The triple is a dual pairing although unlike it is in general not guaranteed to be a dual system. Throughout, unless stated otherwise, all polar sets will be taken with respect to the canonical pairing
Let be a neighborhood of the origin in and let:
be the polar of with respect to the canonical pairing ;
be the polar of with respect to the canonical dual system Note that
A well known fact about polar sets is that
Show that is a -closed subset of Let and suppose that is a net in that converges to in To conclude that it is sufficient (and necessary) to show that for every Because in the scalar field and every value belongs to the closed (in ) subset so too must this net's limit belong to this set. Thus
Show that and then conclude that is a closed subset of both and The inclusion holds because every continuous linear functional is (in particular) a linear functional. For the reverse inclusion let so that which states exactly that the linear functional is bounded on the neighborhood ; thus is a continuous linear functional (that is, ) and so as desired. Using (1) and the fact that the intersection is closed in the subspace topology on the claim about being closed follows.
Conclude that is also a -totally bounded subset of Recall that the topology on is identical to the subspace topology that inherits from This fact, together with (3) and the definition of "totally bounded", implies that is a -totally bounded subset of
Finally, deduce that is a -compact subset of Because is a complete TVS and is a closed (by (2)) and totally bounded (by (4)) subset of it follows that is compact.
If is a normed vector space, then the polar of a neighborhood is closed and norm-bounded in the dual space. In particular, if is the open (or closed) unit ball in then the polar of is the closed unit ball in the continuous dual space of (with the usual dual norm). Consequently, this theorem can be specialized to:
Banach–Alaoglu theorem — If is a normed space then the closed unit ball in the continuous dual space (endowed with its usual operator norm) is compact with respect to the weak-* topology.
When the continuous dual space of is an infinite dimensional normed space then it is impossible for the closed unit ball in to be a compact subset when has its usual norm topology. This is because the unit ball in the norm topology is compact if and only if the space is finite-dimensional (cf. F. Riesz theorem). This theorem is one example of the utility of having different topologies on the same vector space.
It should be cautioned that despite appearances, the Banach–Alaoglu theorem does not imply that the weak-* topology is locally compact. This is because the closed unit ball is only a neighborhood of the origin in the strong topology, but is usually not a neighborhood of the origin in the weak-* topology, as it has empty interior in the weak* topology, unless the space is finite-dimensional. In fact, it is a result of Weil that all locally compactHausdorff topological vector spaces must be finite-dimensional.
Premiere on product/function spaces, nets, and pointwise convergence
For every real will denote the closed ball of radius centered at and for any
Identification of functions with tuples
The Cartesian product is usually thought of as the set of all -indexed tuples but, since tuples are technically just functions from an indexing set, it can also be identified with the space of all functions having prototype as is now described:
Function Tuple: A function belonging to is identified with its (-indexed) "tuple of values"
Tuple Function: A tuple in is identified with the function defined by ; this function's "tuple of values" is the original tuple
This is the reason why many authors write, often without comment, the equality
and why the Cartesian product is sometimes taken as the definition of the set of maps (or conversely). However, the Cartesian product, being the (categorical) product in the category of sets (which is a type of inverse limit), also comes equipped with associated maps that are known as its (coordinate) projections.
The canonical projection of the Cartesian product at a given point is the function
where under the above identification, sends a function to
Stated in words, for a point and function "plugging into " is the same as "plugging into ".
In particular, suppose that are non-negative real numbers. Then where under the above identification of tuples with functions, is the set of all functions such that for every
If a subset partitions into then the linear bijection
canonically identifies these two Cartesian products; moreover, this map is a homeomorphism when these products are endowed with their product topologies. In terms of function spaces, this bijection could be expressed as
Notation for nets and function composition with nets
A net in is by definition a function from a non-empty directed set Every sequence in which by definition is just a function of the form is also a net. As with sequences, the value of a net at an index is denoted by ; however, for this proof, this value may also be denoted by the usual function parentheses notation Similarly for function composition, if is any function then the net (or sequence) that results from "plugging into " is just the function although this is typically denoted by (or by if is a sequence). In the proofs below, this resulting net may be denoted by any of the following notations
depending on whichever notation is cleanest or most clearly communicates the intended information. In particular, if is continuous and in then the conclusion commonly written as may instead be written as or
Topology
The set is assumed to be endowed with the product topology. It is well known that the product topology is identical to the topology of pointwise convergence. This is because given and a net where and every is an element of then the net converges in the product topology if and only if
for every the net converges in
where because and this happens if and only if
for every the net converges in
Thus converges to in the product topology if and only if it converges to pointwise on
This proof will also use the fact that the topology of pointwise convergence is preserved when passing to topological subspaces. This means, for example, that if for every is some (topological) subspace of then the topology of pointwise convergence (or equivalently, the product topology) on is equal to the subspace topology that the set inherits from And if is closed in for every then is a closed subset of
Characterization of
An important fact used by the proof is that for any real
where denotes the supremum and As a side note, this characterization does not hold if the closed ball is replaced with the open ball (and replacing with the strict inequality will not change this; for counter-examples, consider and the identity map on ).
The essence of the Banach–Alaoglu theorem can be found in the next proposition, from which the Banach–Alaoglu theorem follows. Unlike the Banach–Alaoglu theorem, this proposition does not require the vector space to endowed with any topology.
Proposition[3] — Let be a subset of a vector space over the field (where ) and for every real number endow the closed ball with its usual topology ( need not be endowed with any topology, but has its usual Euclidean topology). Define
Before proving the proposition above, it is first shown how the Banach–Alaoglu theorem follows from it (unlike the proposition, Banach–Alaoglu assumes that is a topological vector space (TVS) and that is a neighborhood of the origin).
Proof that Banach–Alaoglu follows from the proposition above
Assume that is a topological vector space with continuous dual space and that is a neighborhood of the origin. Because is a neighborhood of the origin in it is also an absorbing subset of so for every there exists a real number such that Thus the hypotheses of the above proposition are satisfied, and so the set is therefore compact in the weak-* topology. The proof of the Banach–Alaoglu theorem will be complete once it is shown that [note 2] where recall that was defined as
Proof that Because the conclusion is equivalent to If then which states exactly that the linear functional is bounded on the neighborhood thus is a continuous linear functional (that is, ), as desired.
For any let denote the projection to the th coordinate (as defined above). To prove that it is sufficient (and necessary) to show that for every So fix and let Because it remains to show that Recall that was defined in the proposition's statement as being any positive real number that satisfies (so for example, would be a valid choice for each ), which implies Because is a positive homogeneous function that satisfies
Thus which shows that as desired.
Proof of (2):
The algebraic dual space is always a closed subset of (this is proved in the lemma below for readers who are not familiar with this result). The set
is closed in the product topology on since it is a product of closed subsets of Thus is an intersection of two closed subsets of which proves (2).[note 4]
The conclusion that the set is closed can also be reached by applying the following more general result, this time proved using nets, to the special case and
Observation: If is any set and if is a closed subset of a topological space then is a closed subset of in the topology of pointwise convergence.
Proof of observation: Let and suppose that is a net in that converges pointwise to It remains to show that which by definition means For any because in and every value
November 14, 2023
banach, alaoglu, theorem, functional, analysis, related, branches, mathematics, also, known, alaoglu, theorem, states, that, closed, unit, ball, dual, space, normed, vector, space, compact, weak, topology, common, proof, identifies, unit, ball, with, weak, top. In functional analysis and related branches of mathematics the Banach Alaoglu theorem also known as Alaoglu s theorem states that the closed unit ball of the dual space of a normed vector space is compact in the weak topology 1 A common proof identifies the unit ball with the weak topology as a closed subset of a product of compact sets with the product topology As a consequence of Tychonoff s theorem this product and hence the unit ball within is compact This theorem has applications in physics when one describes the set of states of an algebra of observables namely that any state can be written as a convex linear combination of so called pure states Contents 1 History 2 Statement 2 1 Proof involving duality theory 2 2 Elementary proof 3 Sequential Banach Alaoglu theorem 4 Consequences 4 1 Consequences for normed spaces 4 2 Consequences for Hilbert spaces 5 Relation to the axiom of choice and other statements 6 See also 7 Notes 8 Citations 9 References 10 Further readingHistory editAccording to Lawrence Narici and Edward Beckenstein the Alaoglu theorem is a very important result maybe the most important fact about the weak topology that echos throughout functional analysis 2 In 1912 Helly proved that the unit ball of the continuous dual space of C a b displaystyle C a b nbsp is countably weak compact 3 In 1932 Stefan Banach proved that the closed unit ball in the continuous dual space of any separable normed space is sequentially weak compact Banach only considered sequential compactness 3 The proof for the general case was published in 1940 by the mathematician Leonidas Alaoglu According to Pietsch 2007 there are at least twelve mathematicians who can lay claim to this theorem or an important predecessor to it 2 The Bourbaki Alaoglu theorem is a generalization 4 5 of the original theorem by Bourbaki to dual topologies on locally convex spaces This theorem is also called the Banach Alaoglu theorem or the weak compactness theorem and it is commonly called simply the Alaoglu theorem 2 Statement editSee also Topological vector space Dual space Dual system and Polar set If X displaystyle X nbsp is a vector space over the field K displaystyle mathbb K nbsp then X displaystyle X nbsp will denote the algebraic dual space of X displaystyle X nbsp and these two spaces are henceforth associated with the bilinear evaluation map X X K displaystyle left langle cdot cdot right rangle X times X to mathbb K nbsp defined by x f def f x displaystyle left langle x f right rangle stackrel scriptscriptstyle text def f x nbsp where the triple X X displaystyle left langle X X left langle cdot cdot right rangle right rangle nbsp forms a dual system called the canonical dual system If X displaystyle X nbsp is a topological vector space TVS then its continuous dual space will be denoted by X displaystyle X prime nbsp where X X displaystyle X prime subseteq X nbsp always holds Denote the weak topology on X displaystyle X nbsp by s X X displaystyle sigma left X X right nbsp and denote the weak topology on X displaystyle X prime nbsp by s X X displaystyle sigma left X prime X right nbsp The weak topology is also called the topology of pointwise convergence because given a map f displaystyle f nbsp and a net of maps f f i i I displaystyle f bullet left f i right i in I nbsp the net f displaystyle f bullet nbsp converges to f displaystyle f nbsp in this topology if and only if for every point x displaystyle x nbsp in the domain the net of values f i x i I displaystyle left f i x right i in I nbsp converges to the value f x displaystyle f x nbsp Alaoglu theorem 3 For any topological vector space TVS X displaystyle X nbsp not necessarily Hausdorff or locally convex with continuous dual space X displaystyle X prime nbsp the polarU f X sup u U f u 1 displaystyle U circ left f in X prime sup u in U f u leq 1 right nbsp of any neighborhood U displaystyle U nbsp of origin in X displaystyle X nbsp is compact in the weak topology note 1 s X X displaystyle sigma left X prime X right nbsp on X displaystyle X prime nbsp Moreover U displaystyle U circ nbsp is equal to the polar of U displaystyle U nbsp with respect to the canonical system X X displaystyle left langle X X right rangle nbsp and it is also a compact subset of X s X X displaystyle left X sigma left X X right right nbsp Proof involving duality theory edit Proof Denote by the underlying field of X displaystyle X nbsp by K displaystyle mathbb K nbsp which is either the real numbers R displaystyle mathbb R nbsp or complex numbers C displaystyle mathbb C nbsp This proof will use some of the basic properties that are listed in the articles polar set dual system and continuous linear operator To start the proof some definitions and readily verified results are recalled When X displaystyle X nbsp is endowed with the weak topology s X X displaystyle sigma left X X right nbsp then this Hausdorff locally convex topological vector space is denoted by X s X X displaystyle left X sigma left X X right right nbsp The space X s X X displaystyle left X sigma left X X right right nbsp is always a complete TVS however X s X X displaystyle left X prime sigma left X prime X right right nbsp may fail to be a complete space which is the reason why this proof involves the space X s X X displaystyle left X sigma left X X right right nbsp Specifically this proof will use the fact that a subset of a complete Hausdorff space is compact if and only if it is closed and totally bounded Importantly the subspace topology that X displaystyle X prime nbsp inherits from X s X X displaystyle left X sigma left X X right right nbsp is equal to s X X displaystyle sigma left X prime X right nbsp This can be readily verified by showing that given any f X displaystyle f in X prime nbsp a net in X displaystyle X prime nbsp converges to f displaystyle f nbsp in one of these topologies if and only if it also converges to f displaystyle f nbsp in the other topology the conclusion follows because two topologies are equal if and only if they have the exact same convergent nets The triple X X displaystyle left langle X X prime right rangle nbsp is a dual pairing although unlike X X displaystyle left langle X X right rangle nbsp it is in general not guaranteed to be a dual system Throughout unless stated otherwise all polar sets will be taken with respect to the canonical pairing X X displaystyle left langle X X prime right rangle nbsp Let U displaystyle U nbsp be a neighborhood of the origin in X displaystyle X nbsp and let U f X sup u U f u 1 displaystyle U circ left f in X prime sup u in U f u leq 1 right nbsp be the polar of U displaystyle U nbsp with respect to the canonical pairing X X displaystyle left langle X X prime right rangle nbsp U x X sup f U f x 1 displaystyle U circ circ left x in X sup f in U circ f x leq 1 right nbsp be the bipolar of U displaystyle U nbsp with respect to X X displaystyle left langle X X prime right rangle nbsp U f X sup u U f u 1 displaystyle U left f in X sup u in U f u leq 1 right nbsp be the polar of U displaystyle U nbsp with respect to the canonical dual system X X displaystyle left langle X X right rangle nbsp Note that U U X displaystyle U circ U cap X prime nbsp A well known fact about polar sets is that U U displaystyle U circ circ circ subseteq U circ nbsp Show that U displaystyle U nbsp is a s X X displaystyle sigma left X X right nbsp closed subset of X displaystyle X nbsp Let f X displaystyle f in X nbsp and suppose that f f i i I displaystyle f bullet left f i right i in I nbsp is a net in U displaystyle U nbsp that converges to f displaystyle f nbsp in X s X X displaystyle left X sigma left X X right right nbsp To conclude that f U displaystyle f in U nbsp it is sufficient and necessary to show that f u 1 displaystyle f u leq 1 nbsp for every u U displaystyle u in U nbsp Because f i u f u displaystyle f i u to f u nbsp in the scalar field K displaystyle mathbb K nbsp and every value f i u displaystyle f i u nbsp belongs to the closed in K displaystyle mathbb K nbsp subset s K s 1 displaystyle left s in mathbb K s leq 1 right nbsp so too must this net s limit f u displaystyle f u nbsp belong to this set Thus f u 1 displaystyle f u leq 1 nbsp Show that U U displaystyle U U circ nbsp and then conclude that U displaystyle U circ nbsp is a closed subset of both X s X X displaystyle left X sigma left X X right right nbsp and X s X X displaystyle left X prime sigma left X prime X right right nbsp The inclusion U U displaystyle U circ subseteq U nbsp holds because every continuous linear functional is in particular a linear functional For the reverse inclusion U U displaystyle U subseteq U circ nbsp let f U displaystyle f in U nbsp so that sup u U f u 1 displaystyle sup u in U f u leq 1 nbsp which states exactly that the linear functional f displaystyle f nbsp is bounded on the neighborhood U displaystyle U nbsp thus f displaystyle f nbsp is a continuous linear functional that is f X displaystyle f in X prime nbsp and so f U displaystyle f in U circ nbsp as desired Using 1 and the fact that the intersection U X U X U displaystyle U cap X prime U circ cap X prime U circ nbsp is closed in the subspace topology on X displaystyle X prime nbsp the claim about U displaystyle U circ nbsp being closed follows Show that U displaystyle U circ nbsp is a s X X displaystyle sigma left X prime X right nbsp totally bounded subset of X displaystyle X prime nbsp By the bipolar theorem U U displaystyle U subseteq U circ circ nbsp where because the neighborhood U displaystyle U nbsp is an absorbing subset of X displaystyle X nbsp the same must be true of the set U displaystyle U circ circ nbsp it is possible to prove that this implies that U displaystyle U circ nbsp is a s X X displaystyle sigma left X prime X right nbsp bounded subset of X displaystyle X prime nbsp Because X displaystyle X nbsp distinguishes points of X displaystyle X prime nbsp a subset of X displaystyle X prime nbsp is s X X displaystyle sigma left X prime X right nbsp bounded if and only if it is s X X displaystyle sigma left X prime X right nbsp totally bounded So in particular U displaystyle U circ nbsp is also s X X displaystyle sigma left X prime X right nbsp totally bounded Conclude that U displaystyle U circ nbsp is also a s X X displaystyle sigma left X X right nbsp totally bounded subset of X displaystyle X nbsp Recall that the s X X displaystyle sigma left X prime X right nbsp topology on X displaystyle X prime nbsp is identical to the subspace topology that X displaystyle X prime nbsp inherits from X s X X displaystyle left X sigma left X X right right nbsp This fact together with 3 and the definition of totally bounded implies that U displaystyle U circ nbsp is a s X X displaystyle sigma left X X right nbsp totally bounded subset of X displaystyle X nbsp Finally deduce that U displaystyle U circ nbsp is a s X X displaystyle sigma left X prime X right nbsp compact subset of X displaystyle X prime nbsp Because X s X X displaystyle left X sigma left X X right right nbsp is a complete TVS and U displaystyle U circ nbsp is a closed by 2 and totally bounded by 4 subset of X s X X displaystyle left X sigma left X X right right nbsp it follows that U displaystyle U circ nbsp is compact displaystyle blacksquare nbsp If X displaystyle X nbsp is a normed vector space then the polar of a neighborhood is closed and norm bounded in the dual space In particular if U displaystyle U nbsp is the open or closed unit ball in X displaystyle X nbsp then the polar of U displaystyle U nbsp is the closed unit ball in the continuous dual space X displaystyle X prime nbsp of X displaystyle X nbsp with the usual dual norm Consequently this theorem can be specialized to Banach Alaoglu theorem If X displaystyle X nbsp is a normed space then the closed unit ball in the continuous dual space X displaystyle X prime nbsp endowed with its usual operator norm is compact with respect to the weak topology When the continuous dual space X displaystyle X prime nbsp of X displaystyle X nbsp is an infinite dimensional normed space then it is impossible for the closed unit ball in X displaystyle X prime nbsp to be a compact subset when X displaystyle X prime nbsp has its usual norm topology This is because the unit ball in the norm topology is compact if and only if the space is finite dimensional cf F Riesz theorem This theorem is one example of the utility of having different topologies on the same vector space It should be cautioned that despite appearances the Banach Alaoglu theorem does not imply that the weak topology is locally compact This is because the closed unit ball is only a neighborhood of the origin in the strong topology but is usually not a neighborhood of the origin in the weak topology as it has empty interior in the weak topology unless the space is finite dimensional In fact it is a result of Weil that all locally compact Hausdorff topological vector spaces must be finite dimensional Elementary proof edit The following elementary proof does not utilize duality theory and requires only basic concepts from set theory topology and functional analysis What is needed from topology is a working knowledge of net convergence in topological spaces and familiarity with the fact that a linear functional is continuous if and only if it is bounded on a neighborhood of the origin see the articles on continuous linear functionals and sublinear functionals for details Also required is a proper understanding of the technical details of how the space K X displaystyle mathbb K X nbsp of all functions of the form X K displaystyle X to mathbb K nbsp is identified as the Cartesian product x X K textstyle prod x in X mathbb K nbsp and the relationship between pointwise convergence the product topology and subspace topologies they induce on subsets such as the algebraic dual space X displaystyle X nbsp and products of subspaces such as x X B r x textstyle prod x in X B r x nbsp An explanation of these details is now given for readers who are interested Premiere on product function spaces nets and pointwise convergenceFor every real r displaystyle r nbsp B r def c K c r displaystyle B r stackrel scriptscriptstyle text def c in mathbb K c leq r nbsp will denote the closed ball of radius r displaystyle r nbsp centered at 0 displaystyle 0 nbsp and r U def r u u U displaystyle rU stackrel scriptscriptstyle text def ru u in U nbsp for any U X displaystyle U subseteq X nbsp Identification of functions with tuplesThe Cartesian product x X K textstyle prod x in X mathbb K nbsp is usually thought of as the set of all X displaystyle X nbsp indexed tuples s s x x X displaystyle s bullet left s x right x in X nbsp but since tuples are technically just functions from an indexing set it can also be identified with the space K X displaystyle mathbb K X nbsp of all functions having prototype X K displaystyle X to mathbb K nbsp as is now described Function displaystyle to nbsp Tuple A function s X K displaystyle s X to mathbb K nbsp belonging to K X displaystyle mathbb K X nbsp is identified with its X displaystyle X nbsp indexed tuple of values s def s x x X displaystyle s bullet stackrel scriptscriptstyle text def s x x in X nbsp Tuple displaystyle to nbsp Function A tuple s s x x X displaystyle s bullet left s x right x in X nbsp in x X K textstyle prod x in X mathbb K nbsp is identified with the function s X K displaystyle s X to mathbb K nbsp defined by s x def s x displaystyle s x stackrel scriptscriptstyle text def s x nbsp this function s tuple of values is the original tuple s x x X displaystyle left s x right x in X nbsp This is the reason why many authors write often without comment the equalityK X x X K displaystyle mathbb K X prod x in X mathbb K nbsp and why the Cartesian product x X K textstyle prod x in X mathbb K nbsp is sometimes taken as the definition of the set of maps K X displaystyle mathbb K X nbsp or conversely However the Cartesian product being the categorical product in the category of sets which is a type of inverse limit also comes equipped with associated maps that are known as its coordinate projections The canonical projection of the Cartesian product at a given point z X displaystyle z in X nbsp is the functionPr z x X K K defined by s s x x X s z displaystyle Pr z prod x in X mathbb K to mathbb K quad text defined by quad s bullet left s x right x in X mapsto s z nbsp where under the above identification Pr z displaystyle Pr z nbsp sends a function s X K displaystyle s X to mathbb K nbsp to Pr z s def s z displaystyle Pr z s stackrel scriptscriptstyle text def s z nbsp Stated in words for a point z displaystyle z nbsp and function s displaystyle s nbsp plugging z displaystyle z nbsp into s displaystyle s nbsp is the same as plugging s displaystyle s nbsp into Pr z displaystyle Pr z nbsp In particular suppose that r x x X displaystyle left r x right x in X nbsp are non negative real numbers Then x X B r x x X K K X displaystyle prod x in X B r x subseteq prod x in X mathbb K mathbb K X nbsp where under the above identification of tuples with functions x X B r x displaystyle prod x in X B r x nbsp is the set of all functions s K X displaystyle s in mathbb K X nbsp such that s x B r x displaystyle s x in B r x nbsp for every x X displaystyle x in X nbsp If a subset U X displaystyle U subseteq X nbsp partitions X displaystyle X nbsp into X U X U displaystyle X U cup X setminus U nbsp then the linear bijectionH x X K u U K x X U K f x x X f u u U f x x X U displaystyle begin alignedat 4 H amp amp prod x in X mathbb K amp amp to amp left prod u in U mathbb K right times prod x in X setminus U mathbb K 0 3ex amp amp left f x right x in X amp amp mapsto amp left left f u right u in U left f x right x in X setminus U right end alignedat nbsp canonically identifies these two Cartesian products moreover this map is a homeomorphism when these products are endowed with their product topologies In terms of function spaces this bijection could be expressed as H K X K U K X U f f U f X U displaystyle begin alignedat 4 H amp amp mathbb K X amp amp to amp mathbb K U times mathbb K X setminus U 0 3ex amp amp f amp amp mapsto amp left f big vert U f big vert X setminus U right end alignedat nbsp Notation for nets and function composition with netsA net x x i i I displaystyle x bullet left x i right i in I nbsp in X displaystyle X nbsp is by definition a function x I X displaystyle x bullet I to X nbsp from a non empty directed set I displaystyle I leq nbsp Every sequence in X displaystyle X nbsp which by definition is just a function of the form N X displaystyle mathbb N to X nbsp is also a net As with sequences the value of a net x displaystyle x bullet nbsp at an index i I displaystyle i in I nbsp is denoted by x i displaystyle x i nbsp however for this proof this value x i displaystyle x i nbsp may also be denoted by the usual function parentheses notation x i displaystyle x bullet i nbsp Similarly for function composition if F X Y displaystyle F X to Y nbsp is any function then the net or sequence that results from plugging x displaystyle x bullet nbsp into F displaystyle F nbsp is just the function F x I Y displaystyle F circ x bullet I to Y nbsp although this is typically denoted by F x i i I displaystyle left F left x i right right i in I nbsp or by F x i i 1 displaystyle left F left x i right right i 1 infty nbsp if x displaystyle x bullet nbsp is a sequence In the proofs below this resulting net may be denoted by any of the following notationsF x F x i i I def F x displaystyle F left x bullet right left F left x i right right i in I stackrel scriptscriptstyle text def F circ x bullet nbsp depending on whichever notation is cleanest or most clearly communicates the intended information In particular if F X Y displaystyle F X to Y nbsp is continuous and x x displaystyle x bullet to x nbsp in X displaystyle X nbsp then the conclusion commonly written as F x i i I F x displaystyle left F left x i right right i in I to F x nbsp may instead be written as F x F x displaystyle F left x bullet right to F x nbsp or F x F x displaystyle F circ x bullet to F x nbsp TopologyThe set K X x X K textstyle mathbb K X prod x in X mathbb K nbsp is assumed to be endowed with the product topology It is well known that the product topology is identical to the topology of pointwise convergence This is because given f displaystyle f nbsp and a net f i i I displaystyle left f i right i in I nbsp where f displaystyle f nbsp and every f i displaystyle f i nbsp is an element of K X x X K textstyle mathbb K X prod x in X mathbb K nbsp then the net f i i I f displaystyle left f i right i in I to f nbsp converges in the product topology if and only if for every z X displaystyle z in X nbsp the net Pr z f i i I Pr z f displaystyle Pr z left left f i right i in I right to Pr z f nbsp converges in K displaystyle mathbb K nbsp where because Pr z f f z displaystyle Pr z f f z nbsp and Pr z f i i I def Pr z f i i I f i z i I textstyle Pr z left left f i right i in I right stackrel scriptscriptstyle text def left Pr z left f i right right i in I left f i z right i in I nbsp this happens if and only if for every z X displaystyle z in X nbsp the net f i z i I f z displaystyle left f i z right i in I to f z nbsp converges in K displaystyle mathbb K nbsp Thus f i i I displaystyle left f i right i in I nbsp converges to f displaystyle f nbsp in the product topology if and only if it converges to f displaystyle f nbsp pointwise on X displaystyle X nbsp This proof will also use the fact that the topology of pointwise convergence is preserved when passing to topological subspaces This means for example that if for every x X displaystyle x in X nbsp S x K displaystyle S x subseteq mathbb K nbsp is some topological subspace of K displaystyle mathbb K nbsp then the topology of pointwise convergence or equivalently the product topology on x X S x textstyle prod x in X S x nbsp is equal to the subspace topology that the set x X S x textstyle prod x in X S x nbsp inherits from x X K textstyle prod x in X mathbb K nbsp And if S x displaystyle S x nbsp is closed in K displaystyle mathbb K nbsp for every x X displaystyle x in X nbsp then x X S x textstyle prod x in X S x nbsp is a closed subset of x X K textstyle prod x in X mathbb K nbsp Characterization of sup u U f u r displaystyle sup u in U f u leq r nbsp An important fact used by the proof is that for any real r displaystyle r nbsp sup u U f u r if and only if f U B r displaystyle sup u in U f u leq r qquad text if and only if qquad f U subseteq B r nbsp where sup displaystyle sup nbsp denotes the supremum and f U def f u u U displaystyle f U stackrel scriptscriptstyle text def f u u in U nbsp As a side note this characterization does not hold if the closed ball B r displaystyle B r nbsp is replaced with the open ball c K c lt r displaystyle c in mathbb K c lt r nbsp and replacing sup u U f u r displaystyle sup u in U f u leq r nbsp with the strict inequality sup u U f u lt r displaystyle sup u in U f u lt r nbsp will not change this for counter examples consider X def K displaystyle X stackrel scriptscriptstyle text def mathbb K nbsp and the identity map f def Id displaystyle f stackrel scriptscriptstyle text def operatorname Id nbsp on X displaystyle X nbsp The essence of the Banach Alaoglu theorem can be found in the next proposition from which the Banach Alaoglu theorem follows Unlike the Banach Alaoglu theorem this proposition does not require the vector space X displaystyle X nbsp to endowed with any topology Proposition 3 Let U displaystyle U nbsp be a subset of a vector space X displaystyle X nbsp over the field K displaystyle mathbb K nbsp where K R or K C displaystyle mathbb K mathbb R text or mathbb K mathbb C nbsp and for every real number r displaystyle r nbsp endow the closed ball B r def s K s r textstyle B r stackrel scriptscriptstyle text def s in mathbb K s leq r nbsp with its usual topology X displaystyle X nbsp need not be endowed with any topology but K displaystyle mathbb K nbsp has its usual Euclidean topology DefineU def f X sup u U f u 1 displaystyle U stackrel scriptscriptstyle text def Big f in X sup u in U f u leq 1 Big nbsp If for every x X displaystyle x in X nbsp r x gt 0 displaystyle r x gt 0 nbsp is a real number such that x r x U displaystyle x in r x U nbsp then U displaystyle U nbsp is a closed and compact subspace of the product space x X B r x displaystyle prod x in X B r x nbsp where because this product topology is identical to the topology of pointwise convergence which is also called the weak topology in functional analysis this means that U displaystyle U nbsp is compact in the weak topology or weak compact for short Before proving the proposition above it is first shown how the Banach Alaoglu theorem follows from it unlike the proposition Banach Alaoglu assumes that X displaystyle X nbsp is a topological vector space TVS and that U displaystyle U nbsp is a neighborhood of the origin Proof that Banach Alaoglu follows from the proposition above Assume that X displaystyle X nbsp is a topological vector space with continuous dual space X displaystyle X prime nbsp and that U displaystyle U nbsp is a neighborhood of the origin Because U displaystyle U nbsp is a neighborhood of the origin in X displaystyle X nbsp it is also an absorbing subset of X displaystyle X nbsp so for every x X displaystyle x in X nbsp there exists a real number r x gt 0 displaystyle r x gt 0 nbsp such that x r x U displaystyle x in r x U nbsp Thus the hypotheses of the above proposition are satisfied and so the set U displaystyle U nbsp is therefore compact in the weak topology The proof of the Banach Alaoglu theorem will be complete once it is shown that U U displaystyle U U circ nbsp note 2 where recall that U displaystyle U circ nbsp was defined asU def f X sup u U f u 1 U X displaystyle U circ stackrel scriptscriptstyle text def Big f in X prime sup u in U f u leq 1 Big U cap X prime nbsp Proof that U U displaystyle U circ U nbsp Because U U X displaystyle U circ U cap X prime nbsp the conclusion is equivalent to U X displaystyle U subseteq X prime nbsp If f U displaystyle f in U nbsp then sup u U f u 1 displaystyle sup u in U f u leq 1 nbsp which states exactly that the linear functional f displaystyle f nbsp is bounded on the neighborhood U displaystyle U nbsp thus f displaystyle f nbsp is a continuous linear functional that is f X displaystyle f in X prime nbsp as desired displaystyle blacksquare nbsp Proof of Proposition The product space x X B r x textstyle prod x in X B r x nbsp is compact by Tychonoff s theorem since each closed ball B r x def s K s r x displaystyle B r x stackrel scriptscriptstyle text def s in mathbb K s leq r x nbsp is a Hausdorff note 3 compact space Because a closed subset of a compact space is compact the proof of the proposition will be complete once it is shown thatU def f X sup u U f u 1 f X f U B 1 displaystyle U stackrel scriptscriptstyle text def Big f in X sup u in U f u leq 1 Big left f in X f U subseteq B 1 right nbsp is a closed subset of x X B r x textstyle prod x in X B r x nbsp The following statements guarantee this conclusion U x X B r x displaystyle U subseteq prod x in X B r x nbsp U displaystyle U nbsp is a closed subset of the product space x X K K X displaystyle prod x in X mathbb K mathbb K X nbsp Proof of 1 For any z X displaystyle z in X nbsp let Pr z x X K K textstyle Pr z prod x in X mathbb K to mathbb K nbsp denote the projection to the z displaystyle z nbsp th coordinate as defined above To prove that U x X B r x textstyle U subseteq prod x in X B r x nbsp it is sufficient and necessary to show that Pr x U B r x displaystyle Pr x left U right subseteq B r x nbsp for every x X displaystyle x in X nbsp So fix x X displaystyle x in X nbsp and let f U displaystyle f in U nbsp Because Pr x f f x displaystyle Pr x f f x nbsp it remains to show that f x B r x displaystyle f x in B r x nbsp Recall that r x gt 0 displaystyle r x gt 0 nbsp was defined in the proposition s statement as being any positive real number that satisfies x r x U displaystyle x in r x U nbsp so for example r u 1 displaystyle r u 1 nbsp would be a valid choice for each u U displaystyle u in U nbsp which implies u x def 1 r x x U displaystyle u x stackrel scriptscriptstyle text def frac 1 r x x in U nbsp Because f displaystyle f nbsp is a positive homogeneous function that satisfies sup u U f u 1 displaystyle sup u in U f u leq 1 nbsp 1 r x f x 1 r x f x f 1 r x x f u x sup u U f u 1 displaystyle frac 1 r x f x left frac 1 r x f x right left f left frac 1 r x x right right left f left u x right right leq sup u in U f u leq 1 nbsp Thus f x r x displaystyle f x leq r x nbsp which shows that f x B r x displaystyle f x in B r x nbsp as desired Proof of 2 The algebraic dual space X displaystyle X nbsp is always a closed subset of K X x X K textstyle mathbb K X prod x in X mathbb K nbsp this is proved in the lemma below for readers who are not familiar with this result The setU B 1 def f K X sup u U f u 1 f K X f u B 1 for all u U f x x X x X K f u B 1 for all u U x X C x where C x def B 1 if x U K if x U displaystyle begin alignedat 9 U B 1 amp stackrel scriptscriptstyle text def Big f in mathbb K X sup u in U f u leq 1 Big amp big f in mathbb K X f u in B 1 text for all u in U big amp Big left f x right x in X in prod x in X mathbb K f u in B 1 text for all u in U Big amp prod x in X C x quad text where quad C x stackrel scriptscriptstyle text def begin cases B 1 amp text if x in U mathbb K amp text if x not in U end cases end alignedat nbsp is closed in the product topology on x X K K X displaystyle prod x in X mathbb K mathbb K X nbsp since it is a product of closed subsets of K displaystyle mathbb K nbsp Thus U B 1 X U displaystyle U B 1 cap X U nbsp is an intersection of two closed subsets of K X displaystyle mathbb K X nbsp which proves 2 note 4 displaystyle blacksquare nbsp The conclusion that the set U B 1 f K X f U B 1 displaystyle U B 1 left f in mathbb K X f U subseteq B 1 right nbsp is closed can also be reached by applying the following more general result this time proved using nets to the special case Y K displaystyle Y mathbb K nbsp and B B 1 displaystyle B B 1 nbsp Observation If U X displaystyle U subseteq X nbsp is any set and if B Y displaystyle B subseteq Y nbsp is a closed subset of a topological space Y displaystyle Y nbsp then U B def f Y X f U B displaystyle U B stackrel scriptscriptstyle text def left f in Y X f U subseteq B right nbsp is a closed subset of Y X displaystyle Y X nbsp in the topology of pointwise convergence Proof of observation Let f Y X displaystyle f in Y X nbsp and suppose that f i i I displaystyle left f i right i in I nbsp is a net in U B displaystyle U B nbsp that converges pointwise to f displaystyle f nbsp It remains to show that f U B displaystyle f in U B nbsp which by definition means f U B displaystyle f U subseteq B nbsp For any u U displaystyle u in U nbsp because f i u i I f u displaystyle left f i u right i in I to f u nbsp in Y displaystyle Y nbsp and every value f i u f i, wikipedia, wiki, book, books, library,