problem. The proof of Theorem 3 is simply a much more laborious version of the straight- forward strategy above.
Statement
problem. The proof of Theorem 3 is simply a much more laborious version of the straight- forward strategy above. In particular, it produces an explicit algorithm solving the problem of finding triples of \alphai of given cycle structure satisfying \alpha1 \alpha2 = \alpha3 . Then, using also results of Dvir [3], there is an algorithm to find \alpha1 \alpha2 \alpha3 = g for any given g, yielding an elementary proof of Theorem 2. It is quite surprising that a proof of these theorems was not found in the 1980s: the technology involved is essentially the same. The constant \delta and the lower bound on n for which Theorems 2–3 hold are not explicitly stated in the paper, but they can be recovered easily. They are likely to be outside the realm of practical applications, but also likely to be improvable to that point, modulo a reasonable amount of effective work. Most importantly, the proof suggests that problems of this sort on G = Alt(n) can be solved by elementary means, and they can be turned into explicit algo- rithms. The dream goal would be to determine whether C1 C2 ⊇ C3 (or equiva- lently whether \alpha1 \alpha2 = \alpha3 has solutions with \alphai \in Ci ) for any three given classes. It might prove problematic to reach an exact answer in some extreme cases, for instance when |C3 | is very close to |C1 ||C2 |, but Theorem 3 is still quite modest in comparison to what may reasonably be hoped for. For instance, can we constructively prove that C1 C2 ⊇ C3 whenever |C1 |, |C2 | \geq |G|1−\delta and |C3 | \geq |G|10\delta , say? Or even more, can we do it for |Ci | = |G|ηi with |η1 − η2 | + \delta \leq η3 \leq η1 + η2 − \delta? Can we settle the case G = Alt(n) of Problem 19.12 of the Kourovka Notebook, namely, for every normal subset S \subseteq G and every class C \subseteq G the product SC contains at least as many classes as S? Or can we disprove any of the above by families of explicit counterexamples? One important thing to notice is that it does not really matter what the rest of the strings of elements in the \gammai look like: for any triple of \gammai = (\textasciitilde{}\gammai1 3)(4)(\textasciitilde{}\gammai2 9)(10), the manipulation above would work just as well. Moreover, since by hypothesis the classes are large, the cycles are on average quite long, so there are many elements to play with. In our example, even if the elements 3, 4, 9, 10 have been moved around, making some of them unusable, we might still use 2, 8 to do the following: \beta1 : (4 3 1 2)(8 10 9 5 6 7) \beta1 : (4 3 1 2)(8 10 9 5 6 7) \beta2 : (10 4 1 2)(8 9 3 5 6 7) −\to \beta2 (2 8) : (10 4 1 8 9 3 5 6 7 2) \beta1 \beta2 : (1 10 3 2)(8 4 5 7 9 6) \beta1 \beta2 (2 8) : (1 10 3 8 4 5 7 9 6 2)
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.