On universal cycles for multisets (Q1044893): Difference between revisions

A cyclic sequence of elements in \([n] = \{ 1,2,\ldots,n\}\) is universal for \(t\)- -multisets if each \(t\)-multiset of \([n]\) appears once as a subsequence of \(t\) consecutive elements. If such a sequence exists then its length \(\binom{n+t-1}{t}\) must be divisible by \(n\). The paper under review proves that, for \(t=3\), \(4\) and \(6\), a universal cycle for \(t\)--multisets exists if \(n\) is relatively prime with \(t\) (and not too small). For \(t=3\) this show sthat the necessary divisibility condition is also sufficient, and the paper gives three proofs of this case in the hope to set the ground for further results of this type for larger \(t\).