A famous conjecture of Erd\H{o}s and Hajnal (1969) states that for every integer there is a smallest function such that every graph of chromatic number at least contains a subgraph of chromatic number and girth at least . So far, this has only been proved for by R"odl (1977), with bounded by a tower of height . We exhibit a surprising connection between finding high-chromatic subgraphs of large odd-girth (avoiding short odd cycles) and lower-bounding multicolor Ramsey numbers of odd cycles. Using this connection, we prove that for every odd there is a function growing as a power tower of height such that every graph of chromatic number at least contains a subgraph of chromatic number at least and odd-girth at least . This proves a conjecture of Mohar and Wu (2018), addresses a question of Erd\H{o}s and Hajnal (1975), and for improves R"odl’s bound on to a single-exponential. We extend this to a much more general meta-theorem which applies to many graph parameters: if is the fractional chromatic number, the Hall ratio, or the strict vector chromatic number (Lov'{a}sz-Theta-function of the complement), then for every , every graph with sufficiently large contains a subgraph of odd-girth at least with . The key Ramsey-theoretic ingredient is a new lower bound on Ramsey numbers of odd cycles. For , let . We show that for every fixed , where denotes the -fold iterated logarithm. This yields the first superexponential lower bound on multicolor Ramsey numbers of fixed odd cycles, and extends the recent breakthrough by OpenAI for triangles.

生成讲解博客的中文版本

生成AI智能校验的中文版本