Abstract:We study the set of intersection sizes of a k-dimensional affine subspace and a point set of size m \in [0, 2^n] of the n-dimensional binary affine space AG(n,2). Following the theme of Erdős, Füredi, Rothschild and T. Sós, we partially determine which local densities in k-dimensional affine subspaces are unavoidable in all $m$-element point sets in the n-dimensional affine space. We also show constructions of point sets for which the intersection sizes with $k$-dimensional affine subspaces takes values from a set of a small size compared to 2^k. These are built up from affine subspaces and so-called subspace evasive sets. Meanwhile, we improve the best known upper bounds on subspace evasive sets and apply results concerning the canonical signed-digit (CSD) representation of numbers.
What problem does this paper attempt to address?
This paper mainly explores the problem of whether intersections of a given size can be avoided in the finite affine space \(AG(n,2)\). Specifically, the researchers focus on the set of intersection sizes between a \(k\)-dimensional affine subspace and a point set of size \(m\in[0,2^n]\). The core problem that the paper attempts to solve is to determine which local densities in the \(k\)-dimensional affine subspace are unavoidable for all point sets of size \(m\), that is, whether these intersection sizes will always occur.
### Main research contents:
1. **Unavoidability of intersection sizes**: The paper partially determines which local densities in the \(k\)-dimensional affine subspace are unavoidable among all point sets of size \(m\).
2. **Constructing point sets with specific intersection sizes**: The researchers show how to construct point sets so that the intersection sizes between these point sets and the \(k\)-dimensional affine subspace take values within a relatively small set. These constructions are based on affine subspaces and so - called subspace - avoiding sets.
3. **Improvement of subspace - avoiding sets**: The paper also improves the known upper bounds of the best subspace - avoiding sets and applies the results of the Canonical Signed - Digit (CSD) representation of numbers.
### Key concepts:
- **Unavoidability**: If for all point sets \(S\) of size \(m\), there exists a \(k\)-dimensional affine subspace containing exactly \(t\) points, then the pair \([n,m]\) forces the pair \([k,t]\), or the \(t\)-set is said to be unavoidable in the \(k\)-plane.
- **Subspace - avoiding set**: If there exists a point set \(S\subseteq AG(n,q)\) such that all \(k\)-dimensional affine subspaces contain no more than \(c\) points in \(S\), then \(S\) is called a \((k,c)\)-subspace - avoiding set.
### Main results:
- **Theorem 1.5**: If \(0 < t < 2^k\), then \(\text{Sp}(n;k,t)\) will not contain values \(m < 2^n\) whose number of 1s in the binary representation is less than the number of 1s in the binary representation of \(t\).
- **Theorem 1.6**: For fixed \(k,c\) with \(c\geq k + 1\), when \(n > k\), there exists a \((k,c)\)-subspace - avoiding set whose size is at least \(K\cdot q^{n(1 - k/c)}\), where \(K\) is a constant independent of \(n\).
- **Theorem 1.8**: Suppose \(k >\ell\geq1\) are integers, then there exist constants \(C > 0\) and \(0 < D < 1\) such that for \(m\in[C\cdot 2^{n(1 - 1/(2^{k-\ell}-1))},D\cdot 2^n]\), we have \([n,m]\to[k,2^{k-\ell}]\).
- **Theorem 1.9**: For \(k\geq2\), there exists a constant \(C > 0\) such that for \(m\in[C\cdot 2^{n(1 - 1/(2^{k - 2}))},2^n - C\cdot 2^{n(1 - 1/(2^{k - 2}))}]\), we have \([n,m]\to[k,2^{k - 1}]\).
### Conclusion:
The paper uses a variety of methods and tools, such as probabilistic methods, combinatorial methods, and the canonical symbol representation of numbers, to explore in detail the unavoidability problem of specific intersection sizes in the finite affine space, and provides several important theoretical results and construction methods. These results not only deepen the understanding of this problem, but also provide a basis for further research in related fields.