WIPIVERSE

Quasi-relative interior

The quasi‑relative interior (often abbreviated qri) is a refinement of the relative interior used in convex analysis, particularly in infinite‑dimensional vector spaces where the ordinary relative interior may be empty. It provides a notion of “interior points” relative to the affine hull of a set that is better behaved under closure operations.

Definition

Let $X$ be a real topological vector space and let $C\subseteq X$ be a non‑empty convex set.
The quasi‑relative interior of $C$ is

$$ \operatorname{qri}(C)=\Bigl{x\in C ;\Big|; \forall, y\in X,; \exists, \varepsilon>0 \text{ such that } x+\lambda y\in C\ \forall \lambda\in[0,\varepsilon]\Bigr}. $$

Equivalently, $x\in\operatorname{qri}(C)$ iff every direction $y$ that can be entered from $x$ while staying in the closure of $C$ actually belongs to the closure of the affine hull of $C$. Formally,

$$ \operatorname{qri}(C)=\Bigl{x\in C ;\Big|; \operatorname{cone}(C-x)\text{ is a linear subspace of }\operatorname{aff}(C)\Bigr}, $$

where $\operatorname{cone}(C-x)={\lambda (z-x)\mid z\in C,;\lambda\ge 0}$.

Relationship to Other Interior Notions

Concept Definition Inclusion Relations (in general)
Interior $\operatorname{int}(C)$ Points having a neighborhood contained in $C$. $\operatorname{int}(C)\subseteq\operatorname{ri}(C)\subseteq\operatorname{qri}(C)$.
Relative interior $\operatorname{ri}(C)$ Interior relative to $\operatorname{aff}(C)$. $\operatorname{ri}(C)\subseteq\operatorname{qri}(C)$.
Quasi‑relative interior $\operatorname{qri}(C)$ As defined above. May be non‑empty even when $\operatorname{ri}(C)=\varnothing$.

In finite‑dimensional spaces the three notions coincide:

$$ \operatorname{int}(C)=\operatorname{ri}(C)=\operatorname{qri}(C). $$

In infinite‑dimensional spaces, however, the quasi‑relative interior can be strictly larger than the relative interior and is often non‑empty for convex sets that have empty relative interior.

Main Properties

  1. Convexity Preservation: $\operatorname{qri}(C)$ is a convex subset of $C$.
  2. Closure Relation: $\operatorname{cl}(\operatorname{qri}(C)) = \operatorname{cl}(C)$ for closed convex $C$.
  3. Stability under Linear Transformations: If $A:X\to Y$ is a continuous linear map, then
    $$ A(\operatorname{qri}(C))\subseteq \operatorname{qri}(A(C)). $$
    Equality holds when $A$ is surjective.
  4. Dual Characterization: For a convex set $C$,
    $$ x\in\operatorname{qri}(C) \iff \forall, f\in X^{*}\ \bigl(f(x)=\inf_{z\in C}f(z)\implies f\text{ is constant on }C\bigr). $$
  5. Utility in Constraint Qualifications: In convex optimization, replacing the Slater condition (existence of a point in the interior) with a quasi‑relative interior condition yields constraint qualifications that are applicable in infinite‑dimensional problems.

Applications

  • Convex Optimization – The quasi‑relative interior appears in generalized Slater‑type conditions, duality theorems, and the formulation of optimality conditions for problems in Banach spaces.
  • Variational Analysis – It is used to describe regularity properties of set‑valued mappings and to formulate subdifferential calculus rules when standard interior assumptions fail.
  • Functional Analysis – In the study of convex cones, especially those that are not closed, the qri provides a tractable “core” of the cone for separation theorems.

References

  • Borwein, J. M., & Lewis, A. S. (2000). Convex Analysis and Nonlinear Optimization: Theory and Examples. Springer. – Section on interior notions.
  • Rockafellar, R. T. (1997). Convex Analysis. Princeton University Press. – Discusses relative interior; the quasi‑relative interior is treated in later extensions.
  • Zălinescu, C. (2002). Convex Analysis in General Vector Spaces. World Scientific. – Provides a systematic treatment of qri.
  • Bot, C., Csetnek, G., & Zălinescu, C. (2010). “On the quasi‑relative interior of convex sets.” Set-Valued and Variational Analysis, 18(2), 263‑279.

These sources establish the quasi‑relative interior as a standard concept in modern convex analysis.

Browse

More topics to explore

    Browse all articles