Processing math: 100%

Sunday, April 20, 2014

Product of Algebraic Sets and its Coordinate Ring (15.1.28)

MathJax TeX Test Page Prove that if V and W are algebraic sets, then so is V×W and k[V×W]k[V]kk[W].

Proof: If V=Z(I) then also V=Z(I) where I=I(Z(I)), with the additional helpful property that I(V)=I. We shall assume the same for W=Z(J). Lemma: Let VAn, WAm be algebraic sets. Then I(V×W)=I(V)+I(W). Proof: Clearly holds. Now assume fI(V×W) yet fI(V)+I(W). (Following proof of lemma borrowed from MSE) We can write any polynomial in k[x1,...,xn,y1,...,ym]/(I(V)+I(W)) as ri=1¯figi where fik[x1,...,xn] and gik[y1,...,ym] for all i. Let f as above be minimal such that one of its reduction sums has the least summand terms r of the form above of all the polynomials of its hypothesis. Now, assume fi(a)=0 for all aV and all i; then since I(V)=I we have fiI, hence fI(V)+(I(V)+I(W))=I(V)+I(W), a contradiction. Therefore fk(a)0 for some k and aV, where we may reorder f1=fk. Now by the hypothesis for f, for all bW we have f(a)(b)=(fi(a)gi(b))+t(a)(b)=fi(a)gi(b)=0 (for some negligible tI(V)+I(W)) hence fi(a)gi=pI(W). Since f1(a)0 write g1=pf2(a)g2...fr(a)grf1(a) to write ¯f in strictly fewer summands, a contradiction.

Proceed with the exercise. To see that V×W is an algebraic set, observe V×W=Z(I(V)+I(W)), with being evident and resulting from V=Z(I(V)) and W=Z(I(W)). Now, define a k-bilinear map k[V]×k[W]k[V×W] by (f,g)fg (well-defined since I(V)+I(W)I(V×W)) to induce a surjective group homomorphism Φ : k[V]kk[W]k[V×W]. This is also a ring homomorphism. To prove bijection, it suffices to check that the ring homomorphism Φ : k[x1,...xn,y1,...,ym]k[V]kk[W] given by Φ(xi)=xi1 and Φ(yi)=1yi factors through I(V)+I(W), as then it would be a two-sided inverse to Φ when induced on k[V×W]. Indeed, when fI(V) or gI(W) is a generator of I(V×W)=I(V)+I(W), we see Φ(f)=f1=01=0 and Φ(g)=1g=0. 

No comments:

Post a Comment