Longest subsequence for certain repeated up/down patterns in random permutations avoiding a pattern of length three
Abstract: Let $S_n$ denote the set of permutations of $[n]$ and let $\sigma=\sigma_1\cdots\sigma_n\in S_n$. For a subsequence ${\sigma_{i_j}}{j=1}k$ of ${\sigma_i}{i=1}n$ of length $k\ge2$, construct the ``up/down'' sequence $V_1\cdots V_{k-1}$ defined by $$ V_j=\begin{cases} U,\ \text{if}\ \sigma_{i_j+1}-\sigma_{i_j}>0;\ D,\ \text{if}\ \sigma_{i_j+1}-\sigma_{i_j}<0.\end{cases} $$ Consider now a fixed up/down pattern: $V_1\cdots V_l$, where $l\in\mathbb{N}$ and $V_j\in{U, D},\ j\in[l]$. Given a permutation $\sigma\in S_n$, consider the length of the longest subsequence of $\sigma$ that repeats this pattern. For example, consider $l=3$ and $V_1V_2V_3=UUD$. Then for the permutation $342617985\in S_9$, the length of the longest subsequence that repeats the pattern $UUD$ is 7; it is obtained by 3461798 and 3461785. The above framework includes two well-known cases. The pattern $U$ is the celebrated case of the longest increasing subsequence. The pattern $UD$ (or $DU$) is the case of the longest alternating subsequence. These have been studied both under the uniform distribution on $S_n$ as well as under the uniform distribution on those permutations in $S_n$ which avoid a particular pattern of length three. In this paper, we consider the patterns $UUD$ and $UUUD$ under the uniform distribution on those permutations in $S_n$ which avoid the pattern $132$. We prove that the expected value of the longest increasing subsequence following the pattern $UUD$ is asymptotic to $\frac37n$ and the expected value of the longest increasing subsequence following the pattern $UUUD$ is asymptotic to $\frac4{11}n$. (For $UD$ (alternating subsequences) it is known to be $\frac12n$.) This leads directly to appropriate corresponding results for permutations avoiding any particular pattern of length three.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.