Page 119 - Geometric Modeling and Algebraic Geometry
P. 119

120    T. Beck and J. Schicho
                           tried to make the explanation of the necessary concepts and the description of the
                           algorithm as self-contained as possible.
                              Once and for all let K denote a perfect field, the field of definition, and K an alge-
                           braic closure of K. Further let f ∈ K[x, y] be an absolutely irreducible polynomial,
                           i.e. irreducible in K[x, y].



                           7.2 Toric geometry

                           In this section we introduce just as much of toric geometry as we need in this paper. A
                           good general introduction to toric geometry is [3]. We focus on the fact that smooth
                           toric surfaces are generalizations of the standard surfaces P and P × P .
                                                                                 1
                                                                                     1
                                                                          2
                                                                          K      K   K
                           7.2.1 The projective plane P 2
                                                   K
                                                           v
                                                                 y
                           The projective plane is the set of points (¯ :¯x :¯) subject to the equivalence relation
                            v
                                        v
                                  y
                                                y
                           (¯ :¯x :¯)=(λ¯ : λ¯x : λ¯) for λ  =0. It can be covered by 3 affine planes, which
                           are open subsets, depending on whether ¯  =0, ¯x  =0 or ¯  =0. We can introduce
                                                                          y
                                                            v
                           local coordinates on each of these open subsets:
                                                  ⎧     ¯ x  ¯ y
                                                                             v
                                                  ⎨ (1 :  ¯ v  : )=:(1 : u 1 : v 1 ) if ¯  =0
                                                           ¯ v
                                                           ¯ y
                                                     ¯ v
                                        v
                                              y
                                       (¯ :¯x :¯)=  ( :1: )=:(v 2 :1: u 2 ) if ¯x  =0
                                                     ¯ x
                                                           ¯ x
                                                  ⎩ v   ¯ x
                                                    ( :   :1) =: (u 3 : v 3 :1) if ¯  =0
                                                                             y
                                                     ¯
                                                     ¯ y  ¯ y
                              If both sides are defined, i.e. on the intersection of open subsets, we see that
                                                  v i = u −1  ,u i = v i−1 u −1  .        (7.1)
                                                       i−1          i−1
                           Here we assumed for convenience that indices are cyclically arranged, i.e. u 3 = u 0
                           and v 3 = v 0 .
                              The transformation rules for the local coordinates can also be described using a
                           lattice polygon: Draw an isosceles triangle with vertices in Z as in figure 7.1 left,
                                                                             2
                           label the vertices cyclically from 1 to 3 and attach two minimal direction vectors u i
                           and v i to each vertex. Then we find the relations
                                             v i =(−u i−1 ), u i = v i−1 +(−u i−1 )
                           which correspond to (7.1) when passing from additive to multiplicative writing.
                                                 1
                           7.2.2 The ruled surface P × P 1
                                                 K    K
                           The previous example is no coincidence. A similar analogy holds in case of the ruled
                           surface P × P (if a little care is taken when numbering affine charts).
                                   1
                                       1
                                   K   K
                              P × P can be seen as the set of points (¯u, ¯v, ¯x, ¯y) subject to the equivalence
                               1
                                    1
                               K    K
                                         y
                                                         y
                           relation (¯u, ¯v, ¯x, ¯)=(λ¯u, µ¯v, λ¯x, µ¯) for λµ  =0. It can be covered by 4 affine
   114   115   116   117   118   119   120   121   122   123   124