2
$\begingroup$

Let $A$ be a finite dimensional $K$-algebra (where $K$ is a field) and $M$ a finitely generated $A$-module.

Let $\psi: 0 \rightarrow P_r \rightarrow ... \rightarrow P_0 \rightarrow M \rightarrow 0$ be a complex of $A$-modules such that $P_0 \rightarrow M$ is the projective cover of $M$ and $P_r \rightarrow P_{r-1}$ is injective and where the $P_l$ are projective for $l=0,1,...,r$ and the maps $d_i :P_i \rightarrow P_{i-1}$ are minimal, that is $d_i(X) \neq 0$ for any direct summand $X$ of $P_i$. (so $\psi$ is nearly a minimal projective resolution of $M$, the only thing missing is exactness)

Set $P_{-1}:=M$, then it is well known that the Euler characteristic $\chi(\psi):=\sum\limits_{i=-1}^{r}{dim(P_i)}$ is equal to zero in case $\psi$ is exact.

I remember that there is a converse to that, namely that $\psi$ is exact in case $\chi(\psi)=0$ under some extra conditions on $\psi$. I forgot where I saw that. Maybe someone knows a reference for what I have in mind.

Question: Are the conditions here enough to prove that $\psi$ is exact in case $\chi(\psi)=0$ or what other conditions are needed (is there a reference in case this is well known?)?

$\endgroup$

1 Answer 1

2
$\begingroup$

Without more conditions it's not true.

Take the Nakayama algebra with two simples and indecomposable projectives $$P(1)=\matrix{1\\2\\1}\hspace{1cm}\text{and}\hspace{1cm}P(2)=\matrix{2\\1}$$

Then there is a complex $$0\to P(2)\to P(1)\to P(1)\to P(1)\to P(1)\to\matrix{1\\2}\to 0$$ which satisfies your conditions, is not exact, but has zero Euler characteristic.

$\endgroup$
2
  • $\begingroup$ Thanks, I leave the thread open for some suggestions how one could make this true. Maybe one can hope this is true at least for acyclic quiver algebras, since there the possible summands of the projective terms "get smaller" in higher degrees. So one probably cant take repretitions. $\endgroup$
    – Mare
    Commented Apr 23, 2020 at 13:05
  • 1
    $\begingroup$ @Mare For the quiver $1\to2\to3\to4\to5$ with $\text{rad}^3=0$ consider the complex $0\to P(5)\to P(3)\to P(1)\to S(1)\to0$. $\endgroup$ Commented Apr 23, 2020 at 15:13

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .