Sands' Conjecture in Arc-Colored Tournaments

• Sands' Conjecture is a formulation in tournament theory asserting that every k-edge-colored tournament contains a monochromatic dominating set bounded by a function f(k).
• Its analysis leverages iterative partitioning, fractional domination, and probabilistic methods to establish an upper bound of O(ln(2k)·k^(k+2)) for the domination number.
• The conjecture bridges combinatorial and geometric frameworks by linking transitive colorings, quasi-orders, and geometric covering problems, spurring ongoing research in related areas.

Sands' Conjecture (Erdős–Sands–Sauer–Woodrow Conjecture) explores domination phenomena in arc-colored tournaments, positing the existence of small monochromatic dominating sets dependent only on the number of colors used in the coloring. Formally, it asserts that for any positive integer $k$, every $k$-edge-colored tournament contains a monochromatic dominating set of size bounded by some function $f(k)$. Beyond its combinatorial depth, the conjecture tightly links to the structure theory of transitive colorings, quasi-orders, and geometric covering problems.

1. Formal Statement and Key Definitions

Let $T$ be a tournament, i.e., a complete orientation of the edges of a finite simple graph. A $k$-edge coloring assigns to each arc one of $k$ colors. A monochromatic directed path in $T$ is a directed path with all arcs sharing the same color. A subset $S \subseteq V(T)$ is a monochromatic dominating set (also called a kernel by monochromatic paths) if for every $v \notin S$ there exists $s \in S$ and a monochromatic directed path from $s$ to $v$.

Sands' Conjecture: For every positive integer $k$, there exists an integer $f(k)$ such that every $k$-edge-colored tournament $T$ admits a monochromatic dominating set of size at most $f(k)$ (Bousquet et al., 2017).

Equivalent reformulations leverage the quasi-order structure: a partition of the arc set into $k$ quasi-orders suffices for domination number to be bounded by $f(k)$. This general framework links to domination in multidigraphs and generalizations relevant for stability and independence parameters.

2. Relationship with Transitive Colorings and Special Cases

A $k$-transitive coloring of a tournament is one in which the digraph induced by each color class is transitively oriented. If the Sands' Conjecture holds with bound $f(k)$, then the least domination bound $p(k)$ for $k$-transitive tournaments satisfies $p(k) \le f(k)$ (Pálvölgyi et al., 2013). The $k$-transitive case is thus a special, highly structured instance that can provide lower bounds and intuition for the general conjecture.

Refined analysis for $k$-transitive tournaments reveals connections to geometric set systems and VC-dimension apparatus, allowing near-tight bounds on domination numbers in this context through geometric and combinatorial tools.

3. Main Results and Proof Outline

Bousquet, Lochet, and Thomassé (Bousquet et al., 2017) establish:

$\gamma(T)\;=\;O\bigl(\ln(2k)\;k^{\,k+2}\bigr)$

where $\gamma(T)$ is the domination number for any complete multidigraph whose arcs are partitioned into $k$ quasi-orders. Thus, every $k$-edge-colored tournament has a monochromatic dominating set of size $O\bigl(\ln(2k)\;k^{\,k+2}\bigr)$.

Key proof steps involve:

1. Iteratively partitioning the vertex set using fractional domination and probabilistic arguments, ensuring each cell is sufficiently dense in its quasi-order.
2. Applying a two-step domination lemma to produce small dominating sets in dense quasi-orders.
3. Using the pigeonhole principle to manage combinatorial explosion in iterated partitions, bounding total domination number.
4. Recursion over $k$ levels guarantees the total size remains within the stated $O(\ln(2k)k^{k+2})$ bound.

Their approach also extends to digraphs with bounded independence number, showing that the union of $f(k, \alpha)$ stable sets can dominate any multidigraph with maximal stable set size $\alpha$.

4. Geometric and Combinatorial Connections

In the $k$-transitive setting, tournament domination translates to geometric covering problems. For example, the $d$-dimensional box-cover number $g(d)$—the minimal number of $X$-boxes needed to cover any finite $X \subset \mathbb{R}^d$—is closely related:

$g(d) = O(2^{2^{d-1} d \log d})$

improving on prior bounds by linking VC-dimension of the associated hypergraph with domination properties in coordinate tournaments (Pálvölgyi et al., 2013). This is achieved via combinatorial decompositions (e.g., separating into dictatorships, majority, and parity tournaments) and probabilistic shattering methods. In $d=3$, the best known bound was improved to $g(3) \leq 64$.

The interplay between VC-dimension, fractional covers, and hypergraph transversals is central. For any tournament, $\tau^*(H) < 2$ for fractional transversal number, and

$$\tau(H) = O(\mathrm{VC}(H) \,\tau^*(H) \, \log \tau^*(H))$$

where $H$ is the associated covering hypergraph.

5. Open Problems and Further Directions

Several core questions remain unresolved:

• Determining the exact value of $f(3)$. It is unknown whether $f(3) = 3$.
• Pinning down $p(k)$ in the $k$-transitive setting, with particular interest in the domination number of parity tournaments in low dimensions.
• Improving asymptotic upper and lower bounds. For the general setting, current bounds are of tower-exponential type, with simple constructions showing $g(1)=1$, $g(2)=2$, and $g(k)\ge k$.
• The conjecture for multidigraphs seeks not just domination by vertices, but decompositions into stable sets; this remains open in general, though fractional versions exist when stability number is bounded.
• Connections to geometric transversality and box covering suggest that advances in combinatorial geometry could inform future progress.

6. Context, Impact, and Related Problems

Sands' Conjecture, through its general framework on tournaments and colorings, connects to several pillars:

• The Stable Marriage Theorem can be viewed as a statement on the existence of kernels in a 2-layer quasi-order.
• Domination in arc-colored tournaments serves as a combinatorial abstraction for reachability and influence problems on complex networks.
• The geometric analogues influence discrete geometry, particularly in covering and VC-dimension theory.

The conjecture's sharpest possible constraints and extremal structures remain open, and improvements demand refined domination lemmas or more efficient combinatorial decompositions. Its proof techniques (fractional domination, VC-dimension, partition chains) have influenced related lines in extremal combinatorics, theoretical computer science (especially online coloring and dominating set algorithms), and geometric covering theory (Pálvölgyi et al., 2013, Bousquet et al., 2017).

The continued study and refinement of Sands' Conjecture are integral to progress in the understanding of colored tournaments, quasi-order domination, and their combinatorial-geometric interrelations.