Searcharxiv⌕ Search

arXiv · 2611.22032

FunSearch and Structured SAT Search for Weak Schur Partitions

Abstract

The weak Schur problem asks how large an interval of consecutive integers can be partitioned into a fixed number of classes with no class containing a monochromatic solution to $a+b=c$ with $a\neq b$. We investigate whether FunSearch, an evolutionary program-search method guided by a large language model, can discover effective construction strategies for this classical combinatorial problem. In our adaptation, FunSearch evolves a priority function used by a fixed greedy solver, while a separate validity check enforces the weakly sum-free condition. This approach matches the known exact values for two, three, and four classes and reaches $WS(5)\geq195$. We then explore a complementary SAT-based approach. Guided by structural information from a 207-element construction communicated by Fred Rowley (Rowley, 2026), our SAT search produces and verifies a distinct valid five-class partition of $\{1,\ldots,207\}$. Thus the strongest construction obtained in this project reaches $WS(5)\geq207$, eleven integers beyond the best published lower bound of 196. The contrast between the two approaches also exposes a limitation of our FunSearch formulation: its greedy solver makes local, irreversible decisions, whereas SAT can enforce global structure across the whole partition.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Rebecca Vicenzi, Pablo Christofferson, Caleb Hartman, Jingxuan Feng, Suleman Khan, Omar M. Elshinawy, Eric Ramos. 2026-11-17. FunSearch and Structured SAT Search for Weak Schur Partitions. https://searcharxiv.com/abs/2611.22032

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Complexity Framework For Forbidden Subgraphs V: Beyond Simple Graphs

We investigate how allowing self-loops and multi-edges (i.e. moving from simple graphs to pseudographs) impacts the complexity of graph problems on finitely defined monotone graph classes. We do this by extending the C123-framework beyond simple graphs. We use several testbed problems to illustrate some of the new and varied behaviours introduced by the transition to pseudographs. We show that Pseudograph Matching Cut and Pseudograph d-Cut are C123-problems. Consequently, for every finite set HH of simple graphs, we obtain a complete complexity classification for these problems on HH-subgraph-free graphs. This sharply contrasts with their simple-graph counterparts, which are not C123-problems and their complexity remains open even for classes defined by a single forbidden subgraph H. We also show that Pseudograph Stable Cut is not a C123-problem. Nevertheless, for graph classes defined by a single forbidden subgraph, the problem mirrors the C123 classification; this correspondence already fails for sets of size 2. This gives the first example of a problem that aligns with C123-problems for a single forbidden subgraph but not for larger sets. Motivated by the interpretation of Stable Cut as a surjective homomorphism to a partially reflexive path, we also study surjective homomorphisms to partially reflexive cycles on k vertices with one irreflexive vertex. These problems share structural similarities with Pseudograph Stable Cut. We prove that for k at least 6, they are NP-complete on triangle-free pseudographs. However, unlike Pseudograph Stable Cut, they are polynomial-time solvable on all graph classes with bounded-degree.

math.CO↗

A polynomial-time algorithm to compute the co-boxicity of block-structured graphs

The boxicity of a graph $G$ is the minimum dimension $d$ that admits a representation of $G$ as the intersection graph of a family of axis-parallel boxes in $\mathbb{R}^d$. Boxicity has applications in ecology and operations research, where it models competition for shared resources. Computing boxicity is an NP-hard problem. Despite considerable theoretical interest, few graph classes are known for which the boxicity can be computed in polynomial time. Following the approach of Cozzens and Roberts (1983), we study directly the boxicity of graph complements, which we call co-boxicity. In this setting, the geometric intersection problem in $\mathbb{R}^d$ becomes a graph cover problem involving only complements of interval graphs. We give a polynomial-time algorithm that computes the co-boxicity of $G$ from two pieces of information about each block: its co-boxicity, and how the co-boxicity changes when certain subsets of its vertices become cut-vertices. Using this algorithm, we obtain polynomial-time algorithms for the co-boxicity of trees, graphs with bounded block order, block graphs, and cactus graphs. More generally, our method implies the union of these results: for any fixed $k \geq 3$, there is a polynomial-time algorithm to compute the co-boxicity of $G$ whenever every block of $G$ is complete, a cycle, or has at most $k$ vertices.

math.CO↗

Seaweed algebras with restricted part sizes

Seaweed algebras are a class of Lie algebras that are naturally characterized by a pair of compositions, which in turn are represented visually as planar graphs called meanders. These meanders provide a straightforward method for computing the index of the associated algebra. The goal of this paper is to enumerate those seaweed algebras with a fixed index and whose associated compositions have restricted part sizes. In particular, we enumerate those with composition part sizes from so-called acyclic sets. We also establish a bijection between sets of indecomposable seaweed algebras with meanders with certain restricted part sizes and sets of permutations with restricted displacements. In certain cases, the index of the algebra can be determined by a simple statistic on the permutation.

math.CO↗