Conference topic
This is from a problem session at the 2016 West Coast Algebraic Topology Summer School (WCATSS) at The University of Oregon. Thanks to Tyler Lawson for explaining the material.
Definition: Affine schemes are the category Ringop. An object R∈Ring becomes an object Spec(R) in affine schemes, and a ring map R→S becomes a map Spec(S)→Spec(R), where Spec denotes the set of prime ideals.
We try to think of Spec(R) as a geometrical object.
Example: Let k be a field and consider the ring
R=k[x1,…,xn]/(f1(x1,…,xn),…,fr(x1,…,xn)).
Spec(R) is supposed to be a substitute for the set of solutions to a system of equations
f1(x1,…,xn)=0,⋮fr(x1,…,xn)=0.
The scheme Spec(R) has a more precise definition. It consists of a set, a topology, and a sheaf.
1. Set: The underlying set of the scheme Spec(R) is the set of prime ideals of R. For example:
Definition: Affine schemes are the category Ringop. An object R∈Ring becomes an object Spec(R) in affine schemes, and a ring map R→S becomes a map Spec(S)→Spec(R), where Spec denotes the set of prime ideals.
We try to think of Spec(R) as a geometrical object.
Example: Let k be a field and consider the ring
R=k[x1,…,xn]/(f1(x1,…,xn),…,fr(x1,…,xn)).
Spec(R) is supposed to be a substitute for the set of solutions to a system of equations
f1(x1,…,xn)=0,⋮fr(x1,…,xn)=0.
The scheme Spec(R) has a more precise definition. It consists of a set, a topology, and a sheaf.
1. Set: The underlying set of the scheme Spec(R) is the set of prime ideals of R. For example:
- if R=C[x], then the prime ideals are (x−α) and (0);
- if R=C[x,y], then the prime ideals are (x−α,y−β), irreducible polynomials (f(x,y)), and (0).
N⋃n=1V(In)=V(N⋂n=1In)and⋂α∈IV(Iα)=V(∑α∈AIA).
Geometrically, the closed sets are sets of points where one or more identities (like f(x)=0) can hold. For example, if R=C[x], then we have three different closed set types: Spec(C[x]), ∅, or a finite union of (x−α1,…,x−αn). Solutions to equations can be one of the following types below.
3. Sheaf: Let X be a set with a topology. OX is the sheaf for which:
Say R is our ring, Spec(R) our set of primes, and we have some open set U⊆Spec(R). We like to think of it in the following way:
S−1R={[fs] : f∈R,s∈S},
for which OX(U)=S−1R (good enough for today's purposes). Now we have a triple (Spec(R),τ,OX), for τ the Zariski topology, which we call a locally ringed space.
Definition: A scheme is a space X with a topology and a sheaf of rings that is locally isomorphic to Spec(R).
Since the sheaf has the space X and the topology (through the open sets) encoded in it, we may think of a scheme as a special type of sheaf. Also, isomorphism is meant in the category of locally ringed spaces.
Proposition: Morphisms of schemes Spec(R)→Spec(S) are the same as ring maps S→R.
Example: In the Zariski topology, take U⊆Spec(k[x,y]). Locally U looks like it is covered by rings, though that may not be the case globally. Indeed:
Example: Consider projective space P2, where [x:y:z]=[λx:λy:λz]. We may write
P2=U0∪U1∪U2.[1:y:z][x:1:z][x:y:1]Spec(k[y,z])Spec(k[x,z])Spec(k[x,y])
How can we express U0∩U1? This is left as an exercise.
Geometrically, the closed sets are sets of points where one or more identities (like f(x)=0) can hold. For example, if R=C[x], then we have three different closed set types: Spec(C[x]), ∅, or a finite union of (x−α1,…,x−αn). Solutions to equations can be one of the following types below.
3. Sheaf: Let X be a set with a topology. OX is the sheaf for which:
- to each open set U⊆X we get a ring OX(U);
- to each containment V⊆U⊆X of open sets, there exists a restriction map resUV:OX(U)→OX(V);
- the restriction maps are compatible, in the sense that resVW∘resUV=resUW.
Say R is our ring, Spec(R) our set of primes, and we have some open set U⊆Spec(R). We like to think of it in the following way:
- elements of R are functions;
- elements of Spec(R) are points where we can evaluate a function f∈P (or where the function vanishes);
- subsets S⊂R are the sets {f∈R : f only vanishes at points outside U}.
S−1R={[fs] : f∈R,s∈S},
for which OX(U)=S−1R (good enough for today's purposes). Now we have a triple (Spec(R),τ,OX), for τ the Zariski topology, which we call a locally ringed space.
Definition: A scheme is a space X with a topology and a sheaf of rings that is locally isomorphic to Spec(R).
Since the sheaf has the space X and the topology (through the open sets) encoded in it, we may think of a scheme as a special type of sheaf. Also, isomorphism is meant in the category of locally ringed spaces.
Proposition: Morphisms of schemes Spec(R)→Spec(S) are the same as ring maps S→R.
Example: In the Zariski topology, take U⊆Spec(k[x,y]). Locally U looks like it is covered by rings, though that may not be the case globally. Indeed:
Example: Consider projective space P2, where [x:y:z]=[λx:λy:λz]. We may write
P2=U0∪U1∪U2.[1:y:z][x:1:z][x:y:1]Spec(k[y,z])Spec(k[x,z])Spec(k[x,y])
How can we express U0∩U1? This is left as an exercise.
No comments:
Post a Comment