8
$\begingroup$

Let $A$ be a finite dimensional quiver algebra over a field $K$ and let $D(-):=Hom_K(-,K)$ denote the natural duality (assume algebras are connected).

Define $\psi_A:=sup \{ n \geq 1 | D(A)^{\otimes n} \neq 0 \}$ and $W_A:=D(A)^{\otimes \psi_A}$.

Here for an $A$-bimodule $M$, $M^{\otimes n}$ denotes the $n$-fold tensor product of $M$ with itself over $A$.

It seems horribly complicated, but in special cases those definitions might have nice properties. Were they studied before?

Note that Nakayama algebras with a linear quiver and $n$ simple modules are counted by the Catalan nubmers $C_{n-1}$ and thus are in bijection with 321-avoiding permutations on $n-1$-symbols. (Nakayama algebras with a linear quiver are exactly the quotient algebras of the ring of upper triangular matrices over the field $K$ by an admissible ideal)

Here some observations/guesses with the computer for Nakayama algebras with a linear quiver:

  1. $W_A$ is injective.

  2. There are exactly $2^{n-2}$ algebras with $\psi_A = 2$ and all of those algebras seem to have global dimension at most 3.

  3. The generating function of the statistic $A \rightarrow \psi_A$ (http://www.findstat.org/StatisticsDatabase/St001290/) seems to coincide with the generating function on 321-avoiding permutations given by $ \pi \rightarrow f(g(\pi))$, where $g$ is the first fundamental transformation on permutations (http://www.findstat.org/MapsDatabase/Mp00086) and $f$ number of right-to-left minima of a permutation (http://www.findstat.org/StatisticsDatabase/St000991).

Now it looks rather horribly to try to prove those things by direct computations but maybe there is a trick or a nice interpretation of $\psi_A$ and $W_A$ for Nakayama algebras and maybe even more general algebras (maybe QF-3 algebras)?

edit: I found a good reason for 1. and 2. by direct computations with the Nakayama functor applied to the minimal faithful projective-injective module. But maybe there is an argument for 1. that avoids any calculations? 3. remains mysterious.

$\endgroup$

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.