Score:0

How to find k evenly-distributed elements from the set of all n! permutations over n alternatives?

lr flag

Let $C=\{ c_1, c_2, \cdots,c_n \}$ be a set of $n$ alternatives and $T$ be the set of all strict complete orderings on $C$. For any two $t_1$ and $t2$ in $T$, their (Kendal-tau) distance $d(t_1, t_2)$ is defined as the number of pairwise disagreements between $t_1$ and $t_2$.

My Question: How to find $k$ (much smaller than $n!$) different elements from $T$ such that they are "evenly ditributed" in $T$ with respect to this (Kendal-tau) distance $d$?

For example, the k+1 elements $0, 1/k, 2/k, \cdots, (k-1)/k, 1$ are evenly distributed in the interval [0,1].

Hagen von Eitzen avatar
rw flag
$k=n$ cyclic permutations have mutual distance $n$. -- Whenever $n=a+b$ with $a,b>0$, cyclic permutations of the first $a$ and of the last $b$ elements give us $k=ab$ elements such that each has (several) nearest neighbours at distance $\min\{a,b\}$.
lr flag
Thank you so much. So these n cyclic permutations (with mutual distance n) are evenly distributed in the set of all permutations?
lr flag
Can we simply apply the k-medoids algorithm?
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.