2
$\begingroup$

Given a closed-subscheme (corresponding to the homogeneous ideal $I$ inside $K[x_0,...,x_n]$) of a projective scheme $\mathbb{P}^n$, how can one find out the blow up using Macaulay2? Proj(reesAlgebra(I)) gives the blow-up of $I$ along with an affine scheme. Is there any function that we can use? Or we need to do something else?

$\endgroup$

1 Answer 1

5
$\begingroup$

Let $S=k[x_0,\ldots,x_n]$ the polynomial ring and $I=(f_1,\ldots,f_s)$ be the ideal to be blown up. Then the blowup $X$ of $\mathbb{P}^n$ at $I$ is

$$ \mathrm{Proj}(S \oplus I \oplus I^2 \oplus \cdots) = \mathrm{Proj}(T) $$

with $T = S[t I] \subseteq S[t]$ a subring of $S[t]$ ($t$ a new indeterminate). (Proj here in the sense of Hartshorne, Algebraic Geometry, p.160, of an S-Algebra)

Now we have a surjection $\phi:S[T_1,\ldots,T_s] \to S[t I]$ with $\phi:T_i \mapsto t f_i$.

Setting up the map of polynomial rings $\phi:S[T_1,\ldots,T_s] \to S[t]$ is easy in Macaulay2 and it is equally easy to compute $\ker \phi$ (it is an inbuilt operation). So $T$ can be found explicitly as $T = S[T_1,\ldots,T_s]/(\ker\phi)$. To make this the proj of a ring $R$ over $k$ consider the map $$ \psi:R = k[u_{ij}] \to S[T_1,\ldots,T_s], \quad u_{ij} \to x_i T_j $$ where $0 \leqslant i \leqslant n$ and $1 \leqslant j \leqslant s$.

Use this map to form a map $\psi':R \to T$ and compute $J = \ker \psi'$. Then $\mathrm{proj}(R/J)$ is isomorphic to the blowup $X$ from the beginning and is explicitly represented as a closed subscheme of $P^N$ with $N=(n+1) s -1$. (This is an application of the Segre-embedding).

It is possible to extend this discussion to the case of $S=k[x_0,\ldots,x_n]/\mathfrak{a}$ where $\mathfrak{a}$ is an ideal of $k[x_0,\ldots,x_n]$ and to do essentially the same computation with Macaulay2 getting the blow-up of $V(\mathfrak{a})$ at an ideal $\mathfrak{a} + I$.

$\endgroup$
1
  • $\begingroup$ In that case what would be the from proj(R/J) -> P^n? $\endgroup$
    – Anko
    Commented Nov 7, 2018 at 1:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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