arXiv++ Combinatorics

For an integer $n\geq2$, let $G_n$ be the graph on $\{2,\ldots,n\}$ in which distinct integers are adjacent when they have a nontrivial common divisor. Conjecture 448 of Fajtlowicz's \emph{Written on the Wall} asked for a lower bound on the Havel--Hakimi residue $\R(G_n)$. The accompanying historical notes record a stronger lower bound of Erdős and Staton and Staton's conjecture that \(\R(G_n)\sim(ζ(2)-1)n/\log n\). We prove this conjecture and determine the next asymptotic term. If \[ A=\sum_{k=2}^{\infty}\frac{\log k}{k^2(k-1)} =0.3201986326\ldots, \] then \[ \R(G_n)= \bigl(ζ(2)-1\bigr)\frac{n}{\log n} +\bigl(ζ(2)-1-A\bigr)\frac{n}{\log^2 n} +O\!\left(\frac{n}{\log^3 n}\right). \] The lower estimate comes from an exact analysis of the Caro--Wei mass of the prime vertices. For the upper estimate, we construct another realization of the degree sequence of $G_n$: almost every bounded-degree prime vertex is placed in a clique of the order forced by its degree, while degree-preserving switches are absorbed by composite cliques. The proof gives a rare meeting point between prime number asymptotics, degree-sequence algorithms, and automated conjecturing. It also provides a traceable case study in which \textsc{Theo-Conjecture}, an advisor-supervised AI loop, converts registry experiments and counterexamples into a rigorous theorem. We also give an exact Havel--Hakimi defect decomposition. Computation suggests the substantially stronger bound \( \R(G_n)\leq\lceil\CW(G_n)\rceil+2, \) which we state as an open \textsc{Theo-Conjecture} problem.

We investigate machine learning approaches to approximating the \emph{domination number} of graphs, the minimum size of a dominating set. Exact computation of this parameter is NP-hard, restricting classical methods to small instances. We compare two neural paradigms: Convolutional Neural Networks (CNNs), which operate on adjacency matrix representations, and Graph Neural Networks (GNNs), which learn directly from graph structure through message passing. Across 2,000 random graphs with up to 64 vertices, GNNs achieve markedly higher accuracy ($R^2=0.987$, MAE $=0.372$) than CNNs ($R^2=0.955$, MAE $=0.500$). Both models offer substantial speedups over exact solvers, with GNNs delivering more than $200\times$ acceleration while retaining near-perfect fidelity. Our results position GNNs as a practical surrogate for combinatorial graph invariants, with implications for scalable graph optimization and mathematical discovery.

We present four open conjectures in graph theory generated by the automated conjecturing system \texttt{TxGraffiti}. Each conjecture is concise, grounded in natural graph invariants, and empirically validated across hundreds of graphs. Despite extensive effort, these statements remain unresolved--defying both proof and counterexample. They are not only mathematical challenges but creative expressions--born of symbolic pattern recognition and mathematician-defined heuristics, refined through years of human dialogue, and now offered back to the community as collaborative artifacts. These conjectures invite not only formal proof, but also reflection on how machines can evoke wonder, spark curiosity, and contribute to the raw material of discovery. By highlighting these problems, we aim to inspire both human mathematicians and AI systems to engage with them--not only to solve them, but to reflect on what it means when machines participate meaningfully in the creative process of mathematical thought.

A connected forcing set of a graph is a zero forcing set that induces a connected subgraph. In this paper, we introduce and study CF-dense graphs -- graphs in which every vertex belongs to some minimum connected forcing set. We identify several CF-dense graph families and investigate the relationships between CF-density and analogous notions in zero forcing and total forcing. We also characterize CF-dense trees and give a formula for the number of distinct connected forcing sets in trees. Finally, we analyze when CF-density is preserved under graph operations such as Cartesian products, joins, and coronas.

Let $G$ be a graph and $\mathcal{F}$ a family of graphs. Define $α_{\mathcal{F}}(G)$ as the maximum order of any induced subgraph of $G$ that belongs to the family $\mathcal{F}$. For the family $\mathcal{F}$ of graphs with \emph{chromatic number} at most~$k$, we prove that if $G$ is $K_{1,r}$-free, then $α_{\mathcal{F}}(G) \le (r-1)kγ(G)$, where $γ(G)$ is the \emph{domination number}. When $\mathcal{F}$ is the family of empty graphs, this bound simplifies to $α(G) \le 2γ(G)$ for $K_{1,3}$-free (claw-free) graphs, where $α(G)$ is the \emph{independence number} of $G$. For $d$-regular graphs, this is further refined to the bound $α(G) \le 2\left(\frac{d+1}{d+2}\right)γ(G)$, which is tight for $d \in \{2, 3, 4\}$. Using Ramsey theory, we extend this framework to edge-hereditary graph families, showing that for $K_{1,r}$-free graphs, we have $α_{\mathcal{F}}(G) \le r(K_r, \mathcal{F^*})γ(G)$, where $\mathcal{F^*}$ is the set of graphs not in $\mathcal{F}$. Specializing to $K_q$-free graphs, we show $α_{\mathcal{F}}(G) \le (r(K_q, K_r) - 1)γ(G)$. Finally, for the \emph{$k$-independence number} $α_k(G)$, we prove that if $G$ is $K_{1,r}$-free with order $n$ and minimum degree $δ\ge k+1$, \[ α_k(G) \le \left( \frac{(r-1)(k+1)}{δ- k + (r-1)(k+1)} \right) n, \] and this bound is sharp for all parameters.

We prove that the \emph{standard zero forcing number} $Z(G)$ and the \emph{positive semidefinite zero forcing number} $Z_+(G)$ are equal for all claw-free graphs $G$. This result resolves a conjecture proposed by the computer program \emph{TxGraffiti} and highlights a connection between these graph invariants in claw-free structures. As a corollary, we show that a graph $G$ is claw-free if and only if every induced subgraph $H \subseteq G$ satisfies $Z(H) = Z_+(H)$.

This paper introduces the \emph{Optimist}, an autonomous system developed to advance automated conjecture generation in graph theory. Leveraging mixed-integer programming (MIP) and heuristic methods, the \emph{Optimist} generates conjectures that both rediscover established theorems and propose novel inequalities. Through a combination of memory-based computation and agent-like adaptability, the \emph{Optimist} iteratively refines its conjectures by integrating new data, enabling a feedback process with minimal human (\emph{or machine}) intervention. Initial experiments reveal the \emph{Optimist}'s potential to uncover foundational results in graph theory, as well as to produce conjectures of interest for future exploration. This work also outlines the \emph{Optimist}'s evolving integration with a counterpart agent, the \emph{Pessimist} (a human \emph{or machine} agent), to establish a dueling system that will drive fully automated graph theory research.

\emph{TxGraffiti} is a data-driven, heuristic-based computer program developed to automate the process of generating conjectures across various mathematical domains. Since its creation in 2017, \emph{TxGraffiti} has contributed to numerous mathematical publications, particularly in graph theory. In this paper, we present the design and core principles of \emph{TxGraffiti}, including its roots in the original \emph{Graffiti} program, which pioneered the automation of mathematical conjecturing. We describe the data collection process, the generation of plausible conjectures, and methods such as the \emph{Dalmatian} heuristic for filtering out redundant or transitive conjectures. Additionally, we highlight its contributions to the mathematical literature and introduce a new web-based interface that allows users to explore conjectures interactively. While we focus on graph theory, the techniques demonstrated extend to other areas of mathematics.

Let $G$ be a simple graph and let $L(G)$ denote the \emph{line graph} of $G$. A \emph{$p$-independent} set in $G$ is a set of vertices $S \subseteq V(G)$ such that the subgraph induced by $S$ has maximum degree at most $p$. The \emph{$p$-independence number} of $G$, denoted by $α_p(G)$, is the cardinality of a maximum $p$-independent set in $G$. In this paper, and motivated by the recent result that independence number is at most matching number for regular graphs~\cite{CaDaPe2020}, we investigate which values of the non-negative integers $p$, $q$, and $r$ have the property that $α_p(G) \leq α_q(L(G))$ for all r-regular graphs. Triples $(p, q, r)$ having this property are called \emph{valid $α$-triples}. Among the results we prove are: \begin{itemize} \item $(p, q, r)$ is valid $α$-triple for $p \geq 0$, $q \geq 3$ , and $r\geq 2$. \item $(p, q, r)$ is valid $α$-triple for $p \leq q < 3$ and $r\geq 2$. \item $(p, q, r)$ is valid $α$-triple for $p \geq 0$, $q = 2$, and $r$ even. \item $(p, q, r)$ is valid $α$-triple for $p \geq 0$, $q = 2$, and $r$ odd with $r = \max \Big \{ 3, \frac{17(p+1)}{16}\Big \}$. \end{itemize} We also show a close relation between undetermined possible valid $α$-triples, the Linear Aboricity Conjecture, and the Path-Cover Conjecture.

Graph combinatorial optimization problems are widely applicable and notoriously difficult to compute; for example, consider the traveling salesman or facility location problems. In this paper, we explore the feasibility of using convolutional neural networks (CNNs) on graph images to predict the cardinality of combinatorial properties of random graphs and networks. Specifically, we use image representations of modified adjacency matrices of random graphs as training samples for a CNN model to predict the stability number of random graphs; where the stability number is the cardinality of a maximum set of vertices in a graph that contains no pairwise adjacency between vertices. The model and results presented in this study suggest potential for applying deep learning in combinatorial optimization problems previously not considered by simple deep learning techniques.

\emph{TxGraffiti} is a machine learning and heuristic based artificial intelligence designed to automate the task of conjecturing in mathematics. Since its inception, TxGraffiti has generated many surprising conjectures leading to publication in respectable mathematical journals. In this paper we outline the machine learning and heuristic techniques implemented by TxGraffiti. We also recall its contributions to the mathematical literature and announce a new online version of the program available for anyone curious to explore conjectures in graph theory.

For $k \geq 1$ and a graph $G$ without isolated vertices, a \emph{total (distance) $k$-dominating set} of $G$ is a set of vertices $S \subseteq V(G)$ such that every vertex in $G$ is within distance $k$ to some vertex of $S$ other than itself. The \emph{total (distance) $k$-domination number} of $G$ is the minimum cardinality of a total $k$-dominating set in $G$, and is denoted by $γ_{k}^t(G)$. When $k=1$, the total $k$-domination number reduces to the \emph{total domination number}, written $γ_t(G)$; that is, $γ_t(G) = γ_{1}^t(G)$. This paper shows that several known lower bounds on the total domination number generalize nicely to lower bounds on total (distance) $k$-domination.

In the words of the esteemed mathematician Paul Erdös, the mathematician's task is to \emph{prove and conjecture}. These two processes form the bedrock of all mathematical endeavours, and in the recent years, the mathematical community has increasingly sought the assistance of computers to bolster these tasks. This paper is a testament to that pursuit; it presents a robust framework enabling a computer to automatically generate conjectures - particularly those conjectures that mathematicians might deem substantial and elegant. More specifically, we outline our framework and provide evidence in the mathematical literature demonstrating its use in generating publishable research and surprising mathematics. We suspect our simple description of computer-assisted mathematical conjecturing will catalyze further research into this area and encourage the development of more advanced techniques than the ones presented herein.

In this paper, we showcase the process of using an automated conjecturing program called \emph{TxGraffiti} written and maintained by the second author. We begin by proving a conjecture formulated by \emph{TxGraffiti} that for a claw-free graph $G$, the vertex cover number $β(G)$ is greater than or equal to the zero forcing number $Z(G)$. Our proof of this result is constructive, and yields a polynomial time algorithm to find a zero forcing set with cardinality $β(G)$. We also use the output of \emph{TxGraffiti} to construct several infinite families of claw-free graphs for which $Z(G)=β(G)$. Additionally, inspired by the aforementioned conjecture of \emph{TxGraffiti}, we also prove a more general relation between the zero forcing number and the vertex cover number for any connected graph with maximum degree $Δ\ge 3$, namely that $Z(G)\leq (Δ-2)β(G)$+1.

TxGraffiti is an automated conjecturing program that produces graph theoretic conjectures in the form of conjectured inequalities. This program written and maintained by the second author since 2017 was inspired by the successes of previous automated conjecturing programs including Fajtlowicz's GRAFFITI and DeLaViña's GRAFFITI.pc. In this paper we prove and generalize several conjectures generated by TxGraffiti when it was prompted to conjecture on the \emph{independence number}, the \emph{domination number}, and the \emph{matching number} (and generalizations of each of these graph invariants). Moreover, in several instances we also show the proposed inequalities relating these graph invariants are sharp.

In this paper we study relationships between the \emph{matching number}, written $μ(G)$, and the \emph{independence number}, written $α(G)$. Our first main result is to show \[ α(G) \le μ(G) + |X| - μ(G[N_G[X]]), \] where $X$ is \emph{any} intersection of maximum independent sets in $G$. Our second main result is to show \[ δ(G)α(G) \le Δ(G)μ(G), \] where $δ(G)$ and $Δ(G)$ denote the minimum and maximum vertex degrees of $G$, respectively. These results improve on and generalize known relations between $μ(G)$ and $α(G)$. Further, we also give examples showing these improvements.

A set of vertices $S$ in a simple isolate-free graph $G$ is a semi-total dominating set of $G$ if it is a dominating set of $G$ and every vertex of $S$ is within distance 2 or less with another vertex of $S$. The semi-total domination number of $G$, denoted by $γ_{t2}(G)$, is the minimum cardinality of a semi-total dominating set of $G$. In this paper, we study semi-total domination of Cartesian products of graphs. Our main result establishes that for any graphs $G$ and $H$, $γ_{t2}(G\,\square\, H)\ge \frac{1}{3}γ_{t2}(G)γ_{t2}(H)$.

The zero forcing number of a simple graph, written $Z(G)$, is a NP-hard graph invariant which is the result of the zero forcing color change rule. This graph invariant has been heavily studied by linear algebraists, physicists, and graph theorist. It's broad applicability and interesting combinatorial properties have attracted the attention of many researchers. Of particular interest, is that of bounding the zero forcing number from above. In this paper we show a surprising relation between the zero forcing number of a graph and the independence number of a graph, denoted $α(G)$. Our main theorem states that if $G \ne K_4$ is a connected, cubic, claw-free graph, then $Z(G) \le α(G) + 1$. This improves on best known upper bounds for $Z(G)$, as well as known lower bounds on $α(G)$. As a consequence of this result, if $G \ne K_4$ is a connected, cubic, claw-free graph with order $n$, then $Z(G) \le \frac{2}{5}n + 1$. Additionally, under the hypothesis of our main theorem, we further show $Z(G) \le α'(G)$, where $α'(G)$ denotes the matching number of $G$.

A dynamic coloring of the vertices of a graph $G$ starts with an initial subset $S$ of colored vertices, with all remaining vertices being non-colored. At each discrete time interval, a colored vertex with exactly one non-colored neighbor forces this non-colored neighbor to be colored. The initial set $S$ is called a forcing set of $G$ if, by iteratively applying the forcing process, every vertex in $G$ becomes colored. If the initial set $S$ has the added property that it induces a subgraph of $G$ without isolated vertices, then $S$ is called a total forcing set in $G$. The minimum cardinality of a total forcing set in $G$ is its total forcing number, denoted $F_t(G)$. The path cover number of $G$, denoted $\pc(G)$, is the minimum number of vertex disjoint paths such that every vertex belongs to a path in the cover, while the matching number of $G$, denoted $α'(T)$, is the number of edges in a maximum matching of $G$. Let $T$ be a tree of order at least two. We observe that $\pc(T) + 1 \le F_t(T) \le 2\pc(T)$, and we prove that $F_t(T) \le α'(T) + \pc(T)$. Further, we characterize the extremal trees achieving equality in these bounds.

This article studies the \emph{$k$-forcing number} for oriented graphs, generalizing both the \emph{zero forcing number} for directed graphs and the $k$-forcing number for simple graphs. In particular, given a simple graph $G$, we introduce the maximum (minimum) oriented $k$-forcing number, denoted $\MOF_k(G)$ ($\mof_k(G)$), which is the largest (smallest) $k$-forcing number among all possible orientations of $G$. These new ideas are compared to known graph invariants and it is shown that, among other things, $\mof(G)$ equals the path covering number of $G$ while $\MOF_k(G)$ is greater than or equal to the independence number of $G$ -- with equality holding if $G$ is a tree or if $k$ is at least the maximum degree of $G$. Along the way, we also show that many recent results about $k$-forcing number can be modified for oriented graphs.