Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $N$ be a finite set and $\le$ be a partial order. Can you express the number of linear extensions $L(\le,N)$ of this partial order in terms of $b(k) := \#\{l \in N\ :\ k \le l\}$? My intuition says $$L(\le,N) = \frac{n!}{\prod_{k \in N} \max(b(k)-1,1)}$$ where $n = |N|$, but I'm not quite sure.

share|improve this question
add comment

1 Answer

up vote 2 down vote accepted

Your formula gives the wrong result when $\langle N,\le\rangle$ is a linear order: you get $\frac{n!}{(n-1)!}=n$ instead of $1$. That could be corrected by removing the $-1$ in the denominator to make it simply $$L(\le,N) = \frac{n!}{\prod_{k \in N}b(k)}\;,$$ but this also fails. Let $\langle N,\le\rangle=\langle\wp(\{0,1,2\}),\subseteq\rangle$. Then $b(k)=2^{3-|k|}$ for each $k\in N$, so $$\prod_{k\in N}b(k)=\prod_{S\subseteq\{0,1,2\}}2^{3-|S|}=\prod_{k=0}^32^{k\binom3k}=2^{\sum_{k=0}^3k\binom3k}=2^{12}\;,$$ but $8!=2^7\cdot3^2\cdot5\cdot7$ and therefore isn’t divisible by $2^{12}$.

In fact the problem of counting the linear extensions of a partial order is $\#P$-complete.

Added: I worked too hard: the same problem already arises with $\langle\wp(\{0,1\}),\subseteq\rangle$, for which the formula gives $\frac{4!}{4\cdot2\cdot2\cdot1}=\frac32$. This is at least in the right ballpark, since the correct value is $2$. In my original example the figure given by the formula is $9.84375$, which is way off: I count $48$ linear extensions. And on further investigation I find that even the number of linear extensions of the subset order on an $n$-element set appears not to be known in general.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.