Abstract
In this paper, using group actions, we introduce a new method for constructing partial geometric designs (sometimes referred to as $1\frac{1}{2}$-designs). Using this new method, we construct several infinite families of partial geometric designs by investigating the actions of various linear groups of degree two on certain subsets of $\mathbb{F}_{q}^{2}$. Moreover, by computing the stabilizers of such subsets in various linear groups of degree two, we are also able to construct a new infinite family of balanced incomplete block designs.
Introduction
In combinatorial design theory, the construction of sets and subsets (blocks) with specific intersection properties is essential for applications in error-correcting codes, cryptography, and statistical experimental design. A particularly significant class of structures is the partial geometric design, also known as a $1\frac{1}{2}$-design. These are tactical configurations where the number of flags (point-block pairs) follows a specific rule: the number of blocks containing a pair of points depends solely on whether the points are already in a block together.
These designs are of high theoretical interest because their concurrence matrices possess exactly three eigenvalues. This property helps address long-standing questions regarding the spectral properties of block designs, notably those posed by R.A. Bailey. Traditionally, these designs were constructed using "difference sets" or "difference families," where a group acts upon itself. This paper generalizes this approach, allowing a group to act on an arbitrary set, thereby expanding the potential for discovering new combinatorial structures.
Theoretical Framework and Methodology
The core contribution of this work is a new theoretical framework, formalized in Theorem 1, which generalizes the classical construction of designs via group actions.
The General Construction
The method begins with a set $V$ and a group $G$ that acts transitively upon it. By selecting a specific subset (a "base block") and determining its orbit under the action of $G$, a design is formed. The validity of this design as a partial geometric design depends on the existence of two constants, $\alpha$ and $\beta$, which characterize the intersection patterns between points and blocks. Unlike previous methods, this framework does not require the group to act on itself, allowing for a much broader range of base sets.
Linear Group Actions
To demonstrate the utility of this method, the authors focus on the Special Linear Group $SL_2(q)$ and the General Linear Group $GL_2(q)$ acting on vectors in a two-dimensional space over a finite field $\mathbb{F}_q$. The research investigates specific subsets, including:
- Non-zero vectors where neither coordinate is zero.
- Subsets derived from "twin prime cyclotomy," which relates to the parity of exponents of primitive elements in the field.
By calculating the stabilizers—the elements of the group that leave a specific subset unchanged—the authors are able to derive the exact parameters (points, blocks, and replications) of the resulting designs. Computational verification for these structures was performed using the MAGMA software package.
Key Results
The application of this generalized method yielded several new infinite families of designs:
- Theorem 2 (SL_2(q) Families): This theorem establishes an infinite family of partial geometric designs. For instance, when $q=5$, the method produces a $(24, 16, 10; 96, 87)$ design consisting of 15 blocks. The transitive action of $SL_2(q)$ on non-zero vectors proves to be a fertile ground for creating these highly symmetric structures.
- Theorem 3 (Cyclotomy-Based Designs): By utilizing twin prime cyclotomy subsets, the authors identified another family of designs with $2q$ blocks and specific replication numbers.
- Theorem 4 (New BIBDs): By extending the action to the Special Affine Group ($SA_2(q)$), which incorporates translations alongside linear transformations, the authors constructed a new infinite family of Balanced Incomplete Block Designs (BIBDs). These 2-designs are even more restrictive than partial geometric designs. A concrete example provided is a $(49, 18, 357)$ design with 2744 blocks for $q=7$.
Notably, all designs constructed via this method are "simple," meaning they contain no repeated blocks—a property that is often difficult to ensure in other construction techniques.
Discussion and Future Directions
While this research provides a powerful new tool for design construction, it also highlights areas for further study. Calculating certain parameters, such as triple intersection numbers for more complex subsets, remains a significant computational challenge. Additionally, while the parameters of these designs may match those of known "naive" constructions (such as those using Kronecker products), the authors suspect these new designs are non-isomorphic (structurally distinct). Proving this non-isomorphism for all cases remains an open problem.
Future work may involve scaling this methodology to higher-dimensional matrix groups or exploring higher-order cyclotomic classes to uncover even more complex combinatorial structures.
Summary updated