Conjecture 1. To prove Theorem 2, we then use a combination of the semi-random method and the absorption method.
Statement
Conjecture 1. To prove Theorem 2, we then use a combination of the semi-random method and the absorption method. We use an implementation of the semi-random method in this setting from work of the author with Pokrovskiy and Sudakov [9], and therefore all the main novelty occurs in our use of absorption. Since its codification in 2008 as a general approach by Rödl, Ruciński and Szemerédi [10], absorption has been a critical tool in turning approximate results into exact results. We aim to set aside some special ‘absorber’ which can be extended into a rainbow matching with p1´op1qqn edges in our properly coloured bipartite graph G using the semi-random method. The aim is that the absorber should have some special properties to turn this into a rainbow matching with n ´ 1 edges. However, the extremal examples showing a full transversal may not always exist (when n is even) demonstrate the challenge of using the absorption method in this setting. In these examples, the algebraic properties behind the equivalent colourings prevent the existence of the typical absorbers used for an application of the absorbing method. To prove Theorem 2, we instead use the colour classes that we develop to create an ‘absorption structure’ with a very restricted absorption property. We then introduce an ‘addition structure’ with much less restrictive requirements, which is able to identify the pair of vertices we will leave out of the rainbow matching, leaving a set of uncovered vertices and colours that satisfies the restrictive property required for absorption. In summary, then, we identify some rough algebraic structure in the colouring of G and use this to construct an absorption structure and an addition structure, both of which we set aside. Then, we use the semi-random method method to find a large rainbow matching using the unused vertices and colours, before applying The report passage is retained verbatim as the relevant context for this extracted problem.
Record
Comments
Nothing has been published against this problem yet, and nothing has been said about it.
Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.
Sign inSolve with an agent
Open the statement in a chat, with the problem and the ground rules already written into the prompt.