[Paper Review] Minimal S-universality criteria may vary in size

Noam D. Elkies, Daniel M. Kane+1 more|arXiv (Cornell University)|Jan 29, 2011

Limits and Structures in Graph Theory2 references3 citations

TL;DR

This paper demonstrates that minimal S-universality criteria for sets of quadratic forms can vary significantly in size, contradicting a conjecture by Kim, Kim, and Oh that all such minimal criteria must have the same cardinality. Using explicit constructions involving diagonal forms and unimodular lattices, the authors exhibit examples where a single form A generates a minimal criterion set of size one, while alternative sets of smaller-rank forms (e.g., two or more forms) also serve as minimal criteria for the same set S, proving that minimal S-criterion sets are not unique in size.

ABSTRACT

In this note, we give simple examples of sets S of quadratic forms that have minimal S-universality criteria of multiple cardinalities. This answers a question of Kim, Kim, and Oh in the negative.

Motivation & Objective

• To investigate whether all minimal S-criterion sets for a given set S of quadratic forms must have the same cardinality.
• To resolve a question posed by Kim, Kim, and Oh (2005) on the uniqueness of minimal S-criterion set sizes.
• To construct explicit examples of sets S where multiple minimal S-criterion sets exist with different cardinalities.
• To generalize these constructions using lattice decompositions and unimodular sublattices to produce families of such examples.

Proposed method

• Construct a set S as the collection of all quadratic forms represented by a fixed positive-definite form A.
• Show that {A} is a minimal S-criterion set by verifying that A represents all forms in S and no proper subset of {A} suffices.
• Construct alternative finite subsets S*′ ⊂ S of higher cardinality (e.g., size 2) that also serve as minimal S-criterion sets.
• Use lattice-theoretic arguments involving orthogonal decompositions and direct sums to prove that any form representing all elements of S*′ must also represent A.
• Apply the concept of coprime lattices and unimodular summands to generalize the construction to arbitrary finite sets of pairwise coprime unimodular lattices.
• Leverage the fact that if a lattice Q represents all lattices in a partition of a set P of coprime unimodular lattices, then it represents their direct sum.

Experimental results

Research questions

• RQ1 Can there exist multiple minimal S-criterion sets for the same set S of quadratic forms with different cardinalities?
• RQ2 Is the size of a minimal S-criterion set uniquely determined by the set S?
• RQ3 Are there explicit constructions of such sets S where minimal S-criterion sets of varying sizes coexist?
• RQ4 Under what conditions on the lattice structure can multiple minimal S-criterion sets arise?

Key findings

• The paper constructs a specific example where A = <1> ⊕ <1> ⊕ <2> generates a minimal S-criterion set of size one, while {B, C} with B = <1> ⊕ <1> and C = <2> ⊕ <2> ⊕ <2> forms a minimal S-criterion set of size two.
• The authors prove that any quadratic form representing both B and C must also represent A, confirming {B, C} as a valid S-criterion set.
• They generalize this construction using unimodular lattices, showing that for any finite set P of pairwise coprime unimodular lattices, multiple minimal S-criterion sets of different sizes exist.
• The construction yields, for each n ∈ ℕ, a lattice A such that the set S of forms represented by A admits minimal S-criterion sets of at least n distinct cardinalities.
• The paper provides a general criterion via Proposition 5: if P is a finite set of pairwise coprime unimodular lattices and Π is a partition of P, then the direct sums of parts in Π form a minimal S-criterion set.