Linear and Sublinear Diversities

David Bryant,Paul Tupper

2024-12-10

Abstract:Diversities are an extension of the concept of a metric space, where a non-negative value is assigned to any finite set of points, rather than just pairs. Sometimes, diversity theory closely parallels metric theory; other times it veers off in new directions. Here we explore diversities on Euclidean space, particularly those which are Minkowski linear or sublinear. Many well-known functions in convex analysis turn out to be Minkowski linear or Minkowski sublinear diversities, including diameter, circumradius and mean width. We derive characterizations of these classes. Motivated by classical results in metric geometry, and connections with combinatorial optimization, we examine embeddability of finite diversities. We prove that a finite diversity can be embedded into a linear diversity exactly when it has negative type and that it can be embedded into a sublinear diversity exactly when it corresponds to a generalized circumradius.

Metric Geometry

What problem does this paper attempt to address?

The problems that this paper attempts to solve mainly focus on the following aspects:

1. **Properties and characterizations of linear and sublinear diversities**:
- Linear Diversity refers to a diversity function that satisfies specific conditions, including linearity with respect to Minkowski addition and scale - invariance.
- Sublinear Diversity refers to a diversity function that satisfies similar but slightly relaxed conditions.
- The paper aims to explore and characterize the specific forms of these linear and sublinear diversities and study their behaviors in Euclidean spaces.

2. **Embeddability of classical geometric and combinatorial optimization problems**:
- The author explores whether finite diversities can be embedded into linear or sublinear diversities. This involves the concept of the classical Negative Type Metric and the Generalized Circumradius.
- Specifically, the paper proves that a finite diversity can be embedded into a linear diversity if and only if it is of negative type; and it can be embedded into a sublinear diversity if and only if it can be embedded into a Minkowski diversity.

3. **Applications of support functions**:
- When dealing with convex sets and functions defined on them, the support function is an important tool. The paper studies the properties of the support function in detail and uses it to describe the characteristics of linear and sublinear diversities.
- For example, a linear semi - diversity can be represented by a Borel measure on the sphere, which satisfies certain integral conditions.

4. **Combination of theory and application**:
- The research results not only enrich the diversity theory theoretically but also provide a new perspective for combinatorial optimization algorithms. In particular, the paper shows how to approximately solve combinatorial optimization problems on hypergraphs through diversity embedding.

In summary, the main objective of this paper is to deeply study the properties and characterizations of linear and sublinear diversities in Euclidean spaces and their connections with classical geometric and combinatorial optimization problems. Through these studies, the author hopes to expand and develop the diversity theory and provide new methods and tools for its application in practical problems.

### Key formulas

- **Linear diversity**:
\[
\delta(\lambda A)=\lambda\delta(A),\quad\delta(A + B)=\delta(A)+\delta(B)
\]
- **Sublinear diversity**:
\[
\delta(\lambda A)=\lambda\delta(A),\quad\delta(A + B)\leq\delta(A)+\delta(B)
\]
- **Negative - type diversity condition**:
\[
\sum_{A,B}x_Ax_B\delta(A\cup B)\leq0
\]
for all zero - sum vectors \(x\), and \(x_{\emptyset}=0\).
- **Integral representation of linear semi - diversity**:
\[
\delta(A)=\int_{S^{k - 1}}h_A(x)d\nu(x)
\]
where \(S^{k - 1}\) is the unit sphere, \(h_A(x)\) is the support function of the set \(A\), and \(\nu\) is a Borel measure satisfying \(\int_{S^{k - 1}}xd\nu(x)=0\).

These formulas and conclusions provide a solid mathematical basis for understanding linear and sublinear diversities.