3
$\begingroup$

$\newcommand{\red}{\mathrm{red}}$Let $k$ be an algebraically closed field of characteristic zero and $m$ be a positive integer. Let $R$ be the subring $k[x,xy,xy^2,…,xy^m]$ of the polynomial ring $k[x,y]$, and $B$ be the quotient ring $R/xR$. I am trying to compute the $G$-theory groups of the ring $R$.

So far, I have tried to use the $G$-theory localization sequence induced by the fibration sequence $G(R/xR)\rightarrow G(R)\rightarrow G(R_x)$ to do this computation. By using this localization sequence, it suffices to compute the image of the boundary map $\partial:G_1(R_x)\rightarrow G_0(R/xR)$. I have also computed that $B_{\red}\cong k[t]$. So by devissage, we have $$G_0(R/xR)\cong G_0((R/xR)_{\red})=G_0(B_{\red})\cong G_0(k[t])\cong\mathbb{Z}.$$ The isomorphism $G_0(R/xR)\rightarrow G_0((R/xR)_{\red})$ maps the class $[R/xR]$ in $G_0(R/xR)$ to the class $[B/I]+[I/I^2]+…+[I^{m-1}/I^m]$ in $G_0((R/xR)_{\red})$, where $I$ is the nilradical of $B$. This comes from the filtration $0=I^m\subseteq I^{m-1}\subseteq…\subseteq I\subseteq B$.

My question is, what integer does this class in $G_0((R/xR)_{\red})$ correspond to? How do I compute this integer?

Thank you so much for your kind help.

$\endgroup$
5
  • $\begingroup$ Here's an example when $m=3$: let $S=k[t_0,t_1,t_2,t_3]$ and take $\phi:S\rightarrow R$ with $\phi(t_i)=xy^i$. Then the kernel of $\phi$ is (I think; I didn't try to prove these were all the relations) the ideal $(t_1^2=t_0t_2, t_1t_2=t_0t_3, t_2^2=t_1t_3, t_1^3=t_3t_0, t_2^3=t_3^2t_0)$. So $R/xR \cong S/(t_1^2,t_1t_2,t_2^2=t_1t_3, t_2^3):=S'$. And $S'':=S'_{red}\cong k[t_3]$ as you said. The generator of $G_0(S'_{red})$ is $[S'']$. Here if $I=(t_1,t_2)$ in $S'$ is the nilradical, then $S'/I \cong S''$; $I/I^2\cong S''/t_3S'' \oplus S''$ so $[I/I^2]=[S'']$; $I^3=0$ and $I^2\cong S''$ so ... $\endgroup$
    – Eoin
    Commented Feb 4, 2023 at 4:28
  • $\begingroup$ ... in this case, $[S'/I] + [I/I^2] +[I^2] = 3 [S'']$, if I didn't mess up any of the computations. $\endgroup$
    – Eoin
    Commented Feb 4, 2023 at 4:30
  • 1
    $\begingroup$ Thank you so much for your sample reasoning and calculations. I will try to see how the general case works. $\endgroup$
    – Boris
    Commented Feb 4, 2023 at 14:19
  • $\begingroup$ From the comment above, I hope you have worked out that the integer is just $m$. $\endgroup$
    – Mohan
    Commented Feb 6, 2023 at 22:17
  • $\begingroup$ I have tried to compute the quotients $I/I^2$, $I^2/I^3$ and so on as $B/I$-modules to show that they are all isomorphic to $B/I$ in the general case. But I am stuck in this process. $\endgroup$
    – Boris
    Commented Feb 7, 2023 at 0:31

1 Answer 1

1
$\begingroup$

Here is an overkill answer.

Let $A=k[x,xy^m]\subset R$. Then, as I said in a previous answer for a previous question of yours, $R$ is integral over $A$ and thus it is a finite module over $A$. It is torsion free of rank $m$ as a module over $A$. Thus one has a resolution, $0\to P\to F\to R\to 0$ as $A$-modules, where $F$ is a free module of say rank $n$ over $A$. By Auslander-Buchsbaum, $P$ is projective (over $A$) of rank $n-m$ (in fact free by Seshadri's theorem). Tensoring by $A/xA$, we get an exact sequence $0\to P/xP\to F/xF\to B\to 0$. Since $P/xP$ is free (over $A/xA=k[xy^m]$), we see that $[B]=m\in G_0(A/xA)$. On the other hand, we have the natural map $G_0(B_{red})\to G_0(B)$, which as you noted is an isomorphism. Thus, the map $G_0(B)\to G_0(A/xA)$ is an isomorphism, since the composite $G_0(B_{red})\to G_0(A/xA)$ is. This shows the class of $B$ in $G_0(B_{red})$ is $m$.

$\endgroup$
8
  • $\begingroup$ Hello Mohan, thank you so much for kindly providing these detailed explanations of your answer. I am stuck on some parts of your proof. $\endgroup$
    – Boris
    Commented Feb 7, 2023 at 22:04
  • $\begingroup$ 1. How do you show that R is a finitely generated A-module? I understand that R is an integral A-algebra. But if I am not mistaken, we also need R to be a finite-type A-algebra, which I don’t know how to check in this case. $\endgroup$
    – Boris
    Commented Feb 7, 2023 at 22:06
  • $\begingroup$ 2. Assuming that you set P to be the kernel of the surjective A-module homomorphism $F\rightarrow M$, then how do you show that P is free? Thank you so much for your kind help. $\endgroup$
    – Boris
    Commented Feb 7, 2023 at 22:08
  • $\begingroup$ @Boris $R$ is clearly finite type, since it is generated by $xy, xy^2,\ldots, xy^{m-1}$ over $A$. For the second comment, I quoted two non- trivial theorems, though you can get away with Auslander- Buchsbaum. $\endgroup$
    – Mohan
    Commented Feb 7, 2023 at 22:24
  • $\begingroup$ Hello Mohan, thank you so much for your kind help. I understand that $R$ is an $A$-algebra of finite type now. I just found the statement of the Auslander-Buchsbaum formula. Hopefully I will figure out how this formula implies that $P$ is a free $A$-module. $\endgroup$
    – Boris
    Commented Feb 7, 2023 at 23:31

You must log in to answer this question.

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