A note on the best approximation in spaces of affine functions
Maysam Maysami Sadr

TL;DR
This paper investigates the proximinality of subspaces within spaces of bounded affine functions, providing linear analogs of Mazur's classical result using Fenchel's duality theory.
Contribution
It introduces linear versions of an established approximation result for affine functions, employing Fenchel's duality for proofs.
Findings
Proves proximinality of certain subspaces of bounded affine functions.
Establishes linear analogs of Mazur's classical approximation result.
Utilizes sandwich theorems from Fenchel's duality theory in proofs.
Abstract
The proximinality of certain subspaces of spaces of bounded affine functions is proved. The results presented here are some linear versions of an old result due to Mazur. For the proofs we use some sandwich theorems of Fenchel's duality theory.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Banach Space Theory · Fixed Point Theorems Analysis · Optimization and Variational Analysis
A note on the best approximation in spaces of affine functions
Maysam Maysami Sadr [email protected] Department of Mathematics, Institute for Advanced Studies in Basic Sciences, Zanjan, Iran
Abstract
The proximinality of certain subspaces of spaces of bounded affine functions is proved. The results presented here are some linear versions of an old result due to Mazur. For the proofs we use some sandwich theorems of Fenchel’s duality theory.
MSC 2010. Primary 41A65; Secondary 52A07
Keywords. Best approximation, convex set, affine function, sandwich theorem
1 Introduction
Let be a metric space and be a subspace of . Then is called proximinal in if every element of has a best approximation by elements of i.e. for every there exists such that
[TABLE]
The proximinality problem for linear subspaces of normed (function) spaces [7] has been considered by many authors. One of the early results in this direction is due to Mazur:
Theorem 1.1**.**
Let be compact Hausdorff spaces and be a surjective continuous map. Let denote the Banach spaces of real valued continuous functions on with supremum norm. Then the image of in under the canonical linear map induced by , is proximinal.
Mazur’s proof can be found in the Monograph of Semadeni [6] page 124. It uses the Hahn-Tong Sandwich Theorem [6, Theorem 6.4.4] for existence of continuous functions between upper and lower semicontinuous real valued functions. The Mazur result have been extended for spaces of complex valued functions by Pełczyński [5] and for vector valued functions by Olech [4] and Blatter [1]. There is also a generalization for vector spaces of ‘continuous functions’ on ‘noncommutative spaces’ in terms of C*-algebras [8]. For more details see [7, page 15].
In this short note we show that some analogs of the Mazur result are satisfied for spaces of bounded affine functions. Our proofs are linear versions of the Mazur proof. But instead of the Hahn-Tong Theorem we use some sandwich theorems of Fenchel’s duality theory for existence of affine functions between concave and convex functions.
2 The Results
For a nonempty convex set we denote by the normed space of all bounded affine real valued functions on with supremum norm. If is a compact convex subset of a topological vector space we denote by the closed subspace of containing all continuous affine functions. Let be two (compact) convex sets and be a (continuous) surjective affine map. Then induces an isometric linear isomorphism from () into () defined by . In what follows we identify (resp. ) as a closed subspace of (resp. ).
Theorem 2.1**.**
Let be two convex sets and be a surjective affine map. Then , as a subspace of via , is proximinal.
Proof.
Suppose that . We must show that there exists such that where
[TABLE]
Let where
[TABLE]
For and we have,
[TABLE]
This shows that
[TABLE]
and
[TABLE]
Let be bounded real valued functions on defined by
[TABLE]
and
[TABLE]
Then it is easily verified that is convex and is concave. Also,
[TABLE]
and
[TABLE]
By the Sandwich theorem [2, Corollary 2.4.1] there is an affine function with
[TABLE]
Thus and . Hence, by (1) and (2), . ∎
A continuous version of Theorem 2.1 is as follows.
Theorem 2.2**.**
Let be an arbitrary compact convex set and be a compact convex subset of a Fréchet topological vector space. Let be a surjective continuous affine map. Then , as a subspace of via , is proximinal.
Proof.
Suppose that . Let
[TABLE]
Also, let be as in the proof of Theorem 2.1. Thus, for every , (1), (2), (3), and (4) are satisfied. It follows from [6, Lemma 7.5.5] that and are respectively lower and upper semicontinuous functions. By the Sandwich theorem [3, Theorem 6(2)] of Noll, there exists a continuous affine function satisfying (5). Thus and . The proof is complete. ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] J. Blatter, Grothendieck spaces in approximation theory , Mem. Amer. Math. Soc, 120 (1972), 1–121.
- 2[2] J.M. Borwein, J.D. Vanderwerff, Convex functions: constructions, characterizations and counterexamples , Vol. 109 Cambridge: Cambridge University Press, 2010.
- 3[3] D. Noll, Continuous affine support mappings for convex operators , J. Funct. Anal. 76 (1988), 411–431.
- 4[4] C. Olech, Approximation of set-valued functions by continuous functions , Colloq. Math. 19 (1968), 285–293.
- 5[5] A. Pełczyński, Linear extensions, linear averagings and their applications to linear topological classification of spaces of continuous functions , Dissert. Math. (Rozprawy Mat.) 58, Warszawa, 1968.
- 6[6] Z. Semadeni, Banach spaces of continuous functions , PWN Polish Scientific Publishers, 1971.
- 7[7] I. Singer, The theory of best approximation and functional analysis , Society for industrial and applied mathematics, 1974.
- 8[8] D.W.B. Somerset, The proximinality of the centre of a C*-algebra , J. Approx. Theory 89 , no. 1 (1997), 114–117.
