5
$\begingroup$

Let $R$ be a polynomial ring over a field $k$,: $k[x_{1},..x_{n}]$, $\mathfrak{m}=(x_0,...,x_{n})$ and $M$ be a finitely generated $R$ module.

In a paper of Kodiyahlam, he define the Castelnuovo-Mumford regularity of $M$ to be the least integer number $m$ such that for every $j$ the $j$-th syzygy of $M$ is generated in degree less or equal $m+j$.Then, he conclude that the Castelnuovo-Mumford regularity of $M/\mathfrak{m}M$ is equal to the maximal degree of generator of $M$.

Could you please so me why the CM regularity of $M/\mathfrak{m}M$ is equal to the maximal degree of generator of $M$ ?

If $I$ and $J$ are two finitely generated ideal of $R$ and $J\subseteq I$ then do we have the CM regularity of $J$ is less or equal the CM regularity of $I$ ?

$\endgroup$
3
  • 1
    $\begingroup$ The second question has answer "no." $\endgroup$ Commented Oct 8, 2012 at 13:35
  • $\begingroup$ @Charles Staats : Could you please post a counter-example here ? $\endgroup$
    – Knot
    Commented Oct 8, 2012 at 15:20
  • 5
    $\begingroup$ Try resolving $(x,y)$ and $(x^2,y^2)$ over $R=k[x,y]$. The intuition should in fact be the exact opposite: ideals that are "deeper" in the ring should have larger (i.e. worse) regularity. $\endgroup$ Commented Oct 8, 2012 at 16:32

1 Answer 1

3
$\begingroup$

Notice that $M/\mathfrak{m}M$ is an $R$-module of finite length, so $H^i_{\mathfrak{m} }(M/\mathfrak{m}M) = 0$ for all $i>0$ and $H^0_{\mathfrak{m} }(M/\mathfrak{m}M) = M/\mathfrak{m}M$. Recalling that CM regularity can be computed via local cohomology module (see Brodmann-Sharp: local cohomology), we have $$reg (M/\mathfrak{m}M) = \max \{ end (H^i_{\mathfrak{m} }(M/\mathfrak{m}M)) +i | i= 0,...,n \}.$$ Here, consider a graded $R$-module $N = \oplus_iN_i$, we denote $end(N) = \sup \{i | N_i \neq 0\}$. Therefore $reg (M/\mathfrak{m}M) = end ((M/\mathfrak{m}M))$. $end ((M/\mathfrak{m}M))$ is equal to the maximal degree of generator of $M$ by graded Nakayama lemma.

$\endgroup$
3
  • $\begingroup$ What is $end$ ? $\endgroup$ Commented Oct 9, 2012 at 4:17
  • 1
    $\begingroup$ What do you mean by $\text{end}(H^{i}_{\mathfrak{m}}(M/\mathfrak{m}M)$ ? $\endgroup$
    – Knot
    Commented Oct 9, 2012 at 4:19
  • 1
    $\begingroup$ @ Knot: you said that you read a paper of Kodiyahlam about CM regularity. You should study basic fact of CM regularity before reading this paper. $\endgroup$ Commented Oct 9, 2012 at 5:05

You must log in to answer this question.

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