There is no known polynomial-time algorithm for graph isomorphism testing, but elementary combinatorial "refinement" algorithms seem to be very efficient in practice. Some philosophical justification is provided by a classical theorem of Babai, ErdÅs and Selkow: an extremely simple polynomial-time combinatorial algorithm (variously known as "naïve refinement", "naïve vertex classification", "colour refinement" or the "1-dimensional Weisfeiler-Leman algorithm") yields a so-called canonical labelling scheme for "almost all graphs". More precisely, for a typical outcome of a random graph $G(n,1/2)$, this simple combinatorial algorithm assigns labels to vertices in a way that easily permits isomorphism-testing against any other graph.
We improve the Babai-ErdÅs-Selkow theorem in two directions. First, we consider randomly perturbed graphs, in accordance with the smoothed analysis philosophy of Spielman and Teng: for any graph $G$, naïve refinement becomes effective after a tiny random perturbation to $G$ (specifically, the addition and removal of $O(n\log n)$ random edges). Actually, with a twist on naïve refinement, we show that $O(n)$ random additions and removals suffice. These results significantly improve on previous work of Gaudio-Rácz-Sridhar, and are in certain senses best-possible.
Second, we complete a long line of research on canonical labelling of random graphs: for any $p$ (possibly depending on $n$), we prove that a random graph $G(n,p)$ can typically be canonically labelled in polynomial time. This is most interesting in the extremely sparse regime where $p$ has order of magnitude $c/n$; denser regimes were previously handled by Bollobás, Czajka-Pandurangan, and Linial-Mosheiff. Our proof also provides a description of the automorphism group of a typical outcome of $G(n,p_n)$ (slightly correcting a prediction of Linial-Mosheiff).
그래프 동형성 검사 문제는 알려진 다항식 시간 알고리즘이 없지만, 기본적인 조합 "세분화" 알고리즘은 실제로 매우 효율적입니다. Babai, Erdős 및 Selkow의 고전 정리는 이에 대한 철학적 설명을 제공합니다: 극히 단순한 다항식 시간 조합 알고리즘("순진한 세분화", "순진한 정점 분류", "색상 세분화" 또는 "1차원 Weisfeiler-Leman 알고리즘"이라고 불림)이 "거의 모든 그래프"에 대해 정규 표기 방식을 제공할 수 있습니다.
본 논문은 두 가지 방향에서 Babai-Erdős-Selkow 정리를 개선합니다: 첫째, Spielman과 Teng의 평활 분석 개념에 따라 무작위 섭동 그래프를 고려하고; 둘째, 무작위 그래프 정규 표기에 관한 오랜 연구 방향을 완성합니다.