A Takahashi convexity structure on the Isbell-convex hull of an asymmetrically normed real vector space

Philani Rodney Majozi, Mcedisi Sphiwe Zweni · 2026

Read A Takahashi convexity structure on the Isbell-convex hull of an asymmetrically normed real vector space by Philani Rodney Majozi, Mcedisi Sphiwe Zweni (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-12101.txt.

Introduction

Majozi and Zweni lift Takahashi convexity onto a hull built for spaces with asymmetric norms. Earlier work by Künzi and Yildiz set up convexity ideas for T0-quasi-metric spaces. The hull here carries its own canonical quasi-metric and vector-space operations. On that base the authors define a lifted convexity structure with a mixing rule. They show the result is a convex T0-quasi-metric space. The structure also fits the canonical embedding and W-convexity for minimal function pairs.

Catalog record

CREATORPhilani Rodney Majozi, Mcedisi Sphiwe Zweni
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51996
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.12101

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Takahashi convexity structure on the Isbell-convex hull of an asymmetrically normed real vector space Philani Rodney Majozi, Mcedisi Sphiwe Zweni Abstract Let ( X , ∥ ⋅ ∥ ) (X,\|\cdot\|) be an asymmetrically normed real vector space and let ℰ ( X , ∥ ⋅ ∥ ) \mathcal{E}(X,\|\cdot\|) denote its Isbell-convex (injective) hull viewed as a space of minimal ample function pairs. We introduce a canonical T 0 T_{0} -quasi-metric q ℰ q_{\mathcal{E}} on ℰ ( X , ∥ ⋅ ∥ ) \mathcal{E}(X,\|\cdot\|) of sup-difference type and show that the canonical embedding i : X → ℰ ( X , ∥ ⋅ ∥ ) i:X\to\mathcal{E}(X,\|\cdot\|) is isometric. Using the vector space operations on the hull, we define a barycentric map 𝕎 ( f , g , λ ) = λ f ⊕ ( 1 − λ ) g , f , g ∈ ℰ ( X , ∥ ⋅ ∥ ) , λ ∈ [ 0 , 1 ] , \mathbb{W}(f,g,\lambda)=\lambda f\oplus(1-\lambda)g,\qquad f,g\in\mathcal{E}(X,\|\cdot\|),\ \lambda\in[0,1], and prove that ( ℰ ( X , ∥ ⋅ ∥ ) , q ℰ , 𝕎 ) (\mathcal{E}(X,\|\cdot\|),q_{\mathcal{E}},\mathbb{W}) is a convex T 0 T_{0} -quasi-metric space in the sense of Künzi and Yildiz. For the standard affine convexity on X X we establish the equivariance i ⁡ ( W ⁡ ( x , y , λ ) ) = 𝕎 ⁡ ( i ⁡ ( x ) , i ⁡ ( y ) , λ ) i(W(x,y,\lambda))=\mathbb{W}(i(x),i(y),\lambda) , hence i ⁡ ( X ) i(X) is 𝕎 \mathbb{W} -convex in the hull. We further record stability properties of W W -convex function pairs under hull operations and develop a Chebyshev-center/normal-structure framework on ℰ ( X , ∥ ⋅ ∥ ) \mathcal{E}(X,\|\cdot\|) yielding fixed point theorems for nonexpansive self-maps on bounded, doubly closed, 𝕎 \mathbb{W} -convex subsets of the hull. 1 Introduction Convexity structures in the sense of Takahashi provide a flexible way to transfer classical convex-geometric arguments to settings where no linear structure is available. In his seminal work, Takahashi introduced a convexity map on a metric space (X,d) as a rule W(x,y,\lambda) producing “metric convex combinations” and satisfying the characteristic inequality d\bigl(z,W(x,y,\lambda)\bigr)\leq\lambda\,d(z,x)+(1-\lambda)\,d(z,y),\qquad z\in X,\ \lambda\in[0,1], which leads to a rich theory of convex subsets, segments, and fixed point principles for nonexpansive maps in convex metric spaces [ 22 ] . Later developments sharpened this point of view by adding geometric rigidity (for example, strictness/uniqueness and higher-dimensional barycentric devices) and by linking convexity structures to compactness and fixed point theorems for condensing multifunctions [ 23 ] . A systematic asymmetric counterpart was initiated by Künzi and Yildiz, who extended Takahashi’s theory to T_{0} -quasi-metric spaces by requiring a pair of directed inequalities: one for the quasi-metric q and one for its conjugate q^{t} . This two-sided requirement is essential in the asymmetric setting and leads to a coherent notion of a convex T_{0} -quasi-metric space (X,q,W) , together with useful permanence properties, a double-closure formalism for convex sets, and a refined segment theory in the presence of uniqueness [ 15 ] . In particular, uniqueness of a convexity structure forces strong iteration identities and yields quasi-metric segments admitting explicit isometric parametrisations, which highlights the role of strictness notions in asymmetric convexity [ 15 ] . Injective-envelope ideas in metric geometry also admit a powerful asymmetric analogue through the Isbell-convex hull (Isbell-hull). For di-spaces (that is, T_{0} -quasi-metric spaces), Kemajou, Künzi and Olela Otafudu introduced Isbell-convexity as the appropriate replacement of hyperconvexity and proved the fundamental categorical equivalence: a di-space is Isbell-convex if and only if it is di-injective, that is, it has the nonexpansive extension property [ 11 ] . Moreover, every di-space admits an explicit Isbell-hull built from extremal ample function pairs, uniquely determined up to isometry and containing the original space isometrically via the canonical embedding x\mapsto f_{x} [ 11 ] . When the di-space arises from an asymmetrically normed real vector space (X,\|\cdot\|) , this hull is typically denoted \mathcal{E}(X,\|\cdot\|) . A distinctive feature of the asymmetric linear setting is that the Isbell-hull is not merely a “metric” enlargement: it carries canonical algebraic and order structure. Conradie, Künzi and Olela showed that for an asymmetrically normed real vector space (X,\|\cdot\|) the hull \mathcal{E}(X,\|\cdot\|) admits scalar multiplication and an addition operation \oplus turning \mathcal{E}(X,\|\cdot\|) into a real vector space, and that under the specialization order induced by the hull quasi-metric it becomes a Dedekind complete vector lattice [ 5 ] . From a complementary perspective, Olela Otafudu established that extremal function pairs in the asymmetric hull enjoy strong convexity-type regularity and stability properties, reinforcing the view of \mathcal{E}(X,\|\cdot\|) as a natural convex-analytic setting attached to (X,\|\cdot\|) [ 17 ] . The present paper is conceived as a direct follow-up to [ 19 ] , where W -convexity for real-valued function pairs on a convex asymmetrically normed space (X,\|\cdot\|,W) was introduced and developed. In particular, [ 19 ] proved that whenever W is translation-invariant (and under the natural homogeneity requirement), every minimal pair f\in\mathcal{E}(X,\|\cdot\|) is W -convex on X . That work also linked this convexity-of-pairs mechanism to Chebyshev center techniques, normal structure, and fixed point results for nonexpansive maps in convex T_{0} -quasi-metric spaces, building on [ 15 ] and the classical metric theory [ 22 , 23 ] . In this sense, [ 19 ] establishes an inheritance principle: minimal pairs satisfy Jensen-type inequalities on the base space X , provided the underlying convexity structure behaves well under translations. While [ 19 ] shows that hull elements are convex as functions on X , the convexity map itself still acts on X : the barycentric operation takes values in X , and W -convexity is an inequality for the components f_{1},f_{2} evaluated at points of X . On the other hand, the hull \mathcal{E}(X,\|\cdot\|) is a canonical space in its own right: it is the injective (Isbell-convex) envelope of (X,q_{\|\cdot\|}) [ 11 ] and it carries well-defined algebraic operations [ 5 ] . This leads to the following natural question. Can one endow the hull \mathcal{E}(X,\|\cdot\|) itself with a Takahashi convexity structure, defined on the hull, and compatible with the canonical embedding i:X\to\mathcal{E}(X,\|\cdot\|) and the hull operations? The main purpose of this article is to answer this question affirmatively. Using the extremal-pair description of the hull from [ 11 ] and the vector-space operations on \mathcal{E}(X,\|\cdot\|) from [ 5 ] , we introduce: (i) a canonical T_{0} -quasi-metric q_{\mathcal{E}} on \mathcal{E}(X,\|\cdot\|) of sup-difference type, chosen so that the canonical embedding i(x)=f_{x} is isometric, and (ii) a lifted convexity map \mathbb{W}:\mathcal{E}(X,\|\cdot\|)\times\mathcal{E}(X,\|\cdot\|)\times[0,1]\to\mathcal{E}(X,\|\cdot\|),\qquad\mathbb{W}(f,g,\lambda):=\lambda f\ \oplus\ (1-\lambda)g, defined using the hull operations (\oplus,\cdot) . We then prove that (\mathcal{E}(X,\|\cdot\|),q_{\mathcal{E}},\mathbb{W}) is a convex T_{0} -quasi-metric space in the sense of Künzi and Yildiz [ 15 ] (Theorem 3.11 ). This places a convex geometry on the hull itself, not only convexity inequalities for hull elements viewed as functions on X . A second main result establishes compatibility with the base space. In the linear situation where W is the standard affine convexity on X , we show that the embedding i:X\to\mathcal{E}(X,\|\cdot\|) intertwines the base convexity and the lifted hull convexity: i\bigl(W(x,y,\lambda)\bigr)=\mathbb{W}\bigl(i(x),i(y),\lambda\bigr),\qquad x,y\in X,\ \lambda\in[0,1] (Theorem 3.13 ). In particular, i(X) is \mathbb{W} -convex in \mathcal{E}(X,\|\cdot\|) ,

Read the complete text