1
$\begingroup$

I am asked to calculate the Hilbert polynomial of the twisted cubic curve \begin{equation*} C = \{(s^3 : s^2t : st^2 : t^3); (s:t) \in \mathbb{P}^1 \} \subset \mathbb{P}^3 \end{equation*} and I know it should be $2d-1$. In the examples in my book there are some simple examples where the Hilbert polynomial is calculated by calculating the basis of $S(X)^{(d)}$ , so I also tried to do this, by using that $C$ is the zero-set of \begin{equation*} \{x_0x_2 - {x_1}^2, x_1x_3 - {x_2}^2, x_0x_3 - x_1x_2 \}. \end{equation*} I am trying to calculate a basis for $S(X)^{(d)}$ with $2d-1$ elements by scoring out the elements of the zero-set, but the calculations get quit complicated and I am not sure if this is the right approach.

$\endgroup$
  • 3
    $\begingroup$ possible duplicate of Hilbert polynomial Twisted cubic There is a nice way to do this "by hand" more elementarily though by realizing the coordinate ring as k[s3,s2t,st2,t3], and counting monomials there to get that the Hilbert polynomial is 3d+1 $\endgroup$ – Takumi Murayama May 26 '15 at 12:36
1
$\begingroup$

Your approach is OK till this point, find a Grobner basis then use that to find a standard monomial basis.

$\endgroup$
5
$\begingroup$

First, the coordinate ring of the twisted cubic is $k[s^3,s^2t,st^2,t^3]$, since $[s:t] \mapsto [s^3:s^2t:st^2:t^3]$ is a parametrization of the twisted cubic. Note that in particular, requiring that the ring homomorphism $k[w,x,y,z] \to k[s^3,s^2t,st^2,t^3]$ inducing the isomorphism is in fact graded means that the generators $s^3,s^2t,st^2,t^3$ each live in the grade one piece of the algebra. See Georges Elencwajg's answer to another question for a proof.

It therefore suffices to find the Hilbert polynomial for $k[s^3,s^2t,st^2,t^3]$. But the monomials in this algebra are in one-to-one correspondence with the monomials in $k[s,t]$ living in each graded piece that is a multiple of three. Since the Hilbert polynomial for $k[s,t]$ is $d+1$, the Hilbert polynomial for $k[s^3,s^2t,st^2,t^3]$ is therefore $3d+1$.

I want to point out that the above argument is a nice way to compute other Hilbert polynomials: for example, if $X$ is the image of the Segre embedding $\mathbf{P}^1 \times \mathbf{P}^2 \to \mathbf{P}^5$, the Hilbert polynomial of $X$ is $h_X(d) = (d+1)(d/2 + 2)$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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