Caccetta-Häggkvist Conjecture
Status open high confidence
The Caccetta-Häggkvist conjecture remains open in its full generality, including the well-known case $r = n/3$ (existence of a directed triangle). Post-2013 work has produced only partial results: improvements on the digraph triangle bound, proofs for restricted classes (e.g., small independence number, forbidden subgraphs), counterexamples to a stronger conjecture of Thomassé, and progress on Aharoni's rainbow generalization (now resolved up to a constant factor and up to an additive constant).
Cited literature (9)
-
Disproves Thomassé's strengthening of the Caccetta-Häggkvist conjecture for all girth $g \geq 4$ and proves new lower bounds on directed path lengths in digraphs of given minimum out-degree and girth.
-
partial Improved bounds for the triangle case of Aharoni's rainbow generalization of the Caccetta-Häggkvist conjecture (2022)
Establishes improved triangle-existence bounds in Aharoni's rainbow generalization of the Caccetta-Häggkvist conjecture, e.g. that $(1.1077, 1/3)$ is triangular and $(1, 0.3988)$ is triangular.
-
partial Further approximations for Aharoni's rainbow generalization of the Caccetta-Häggkvist conjecture (2021)
Proves Aharoni's rainbow version of the Caccetta-Häggkvist conjecture holds when each color class has size $\Omega(k)$, improving the previous $\Omega(k \log k)$ requirement and bringing the result to within a constant factor of the conjecture.
-
Shows that Aharoni's rainbow generalization of the Caccetta-Häggkvist conjecture holds up to an additive constant: edge-colored graphs with $n$ vertices and $n$ color classes of size at least $r$ contain a rainbow cycle of length at most $n/r + \alpha_r$.
-
Proves that for edge-colored graphs on $n$ vertices with $n$ color classes (each a matching of size 2 or a triangle) of which at least $\alpha n$ have a given form for $\alpha > 1/2$, the graph contains a rainbow cycle of length $O(\log n)$, contributing to Aharoni's rainbow version of the Caccetta-Häggkvist conjecture.
-
Proves that every locally in-tournament digraph satisfies the Caccetta-Häggkvist conjecture, deducing it from the structural decomposition theorem (Theorem 2.3).
-
partial Improved bounds for the triangle case of Aharoni's rainbow generalization of the Caccetta-Häggkvist conjecture (2023)
Does not prove new bounds for the conjecture directly; instead uses existing approximations for the triangle case as a tool to establish improved triangularity bounds for Aharoni's rainbow generalization, and notes that improvements to those approximations would yield improvements to Theorem 3.
-
Proves the $r=2$ case of Aharoni's rainbow strengthening of this conjecture (Theorem 5), which implies the $r=2$ case of Caccetta-Häggkvist. Also introduces Conjecture 4, a weaker rainbow properly-coloured-cycle condition shown to imply the full Caccetta-Häggkvist conjecture.
-
Shows that the Caccetta-Häggkvist conjecture implies the bipartite analogue Conjecture 1.2; uses CH-approximations to prove 1.2 for $k = 1, 2, 3, 4, 6$, and all $k \geq 224{,}539$.
Reviewer notes. The original Caccetta-Häggkvist conjecture is unresolved, including the central r = n/3 (triangle) case. Most recent activity post-2013 concerns either Aharoni's rainbow generalization or strengthenings/restrictions (small independence number, forbidden subgraphs). The 2024 Cheng-Keevash paper disproves a strictly stronger conjecture of Thomassé but does not affect the original Caccetta-Häggkvist statement. The SIAM page for Devlin/Aharoni-style 'Nonuniform Degrees' rainbow paper returned 403 and was not directly verified, so it is not cited.
Discussion
It is one of the most famous conjectures in graph theory. It has many alternative formulations and lots of work have been done around it. Many interesting conjectures are related to it. See [Sul]. It is in particular implied by a conjecture of Thomassé and Hoàng-Reed Conjecture . The Caccetta-Häggkvist Conjecture is a generalization of an earlier conjecture of Behzad, Chartrand, and Wall, who conjectured it only for diregular digraphs. Caccetta-H äggkvist Conjecture has been proved for $ r\leq \sqrt{n/2} $ by Shen [She1]. For $ r\geq n/2 $ it is trivial. But already for $ r=n/3 $ , it is still open as well as Behzad-Chartrand-Wall Conjecture Conjecture Every simple $ n $ -vertex digraph with minimum outdegree at least $ r/3 $ and minimum indegree at least $ r/3 $ has a cycle with length at most $ 3 $ . This conjecture would be implied by Seymour's Second Neighbourhood Conjecure . Shen [She2] also proved the following approximate version. Theorem Every simple digraph of order $ n $ with minimum outdegree at least $ r $ has a cycle with length at most $ n/r + 73 $ . Bollobás and Scott [BS] proposed a weighted version of the Caccetta-Häggkvist Conjecture. Conjecture Let $ w:E(D) \rightarrow [0,1] $ be a weight function on the arcs of a digraph $ D $ . If $ \sum_{u\in N^-(v)} w(uv) \geq 1 $ and $ \sum_{u\in N^+(v)} w(vu) \geq 1 $ for all $ v\in V(D) $ , then there is a directed cycle in $ D $ of total weight at least 1. They gave a nice proof that there is a directed path of total weight at least 1.
Bibliography
-
[BCW]M. Behzad, G. Chartrand, and C. Wall. On minimal regular digraphs with given girth. Fundamenta Mathematicae, 69:227–231, 1970. -
[BS]B. Bollobás and A. D. Scott, A proof of a conjecture of {B}ondy concerning paths in weighted digraphs. J. Combin. Theory Ser. B, 66:283-292, 1996. -
★
[CH]L. Caccetta and R. Häggkvist. On minimal digraphs with given girth. Congressus Numerantium, XXI, 1978 -
[?][She1J. Shen. On the girth of digraphs. Discrete Math, 211(1-3):167–181, 2000. -
[She2]J. Shen. On the Caccetta-Häggkvist conjecture. Graphs and Combinatorics, 18(3):645–654, 2002. -
[Sul]Blair D. Sullivan: A Summary of Problems and Results related to the Caccetta-Häggkvist Conjecture A Summary of Problems and Results related to the Caccetta-Häggkvist Conjecture