A Virtually Complete Classification of Virtual Complete Intersections - - PowerPoint PPT Presentation

a virtually complete classification of virtual
SMART_READER_LITE
LIVE PREVIEW

A Virtually Complete Classification of Virtual Complete Intersections - - PowerPoint PPT Presentation

Outline Preliminaries Determination of VCIs A Virtually Complete Classification of Virtual Complete Intersections in P 1 P 1 Jiyang Gao, Yutong Li, Amal Mattoo University of Minnesota - Twin Cities REU 2018 1 August 2018 VCIs in P 1 P 1


slide-1
SLIDE 1

Outline Preliminaries Determination of VCIs

A Virtually Complete Classification of Virtual Complete Intersections in P1 × P1

Jiyang Gao, Yutong Li, Amal Mattoo

University of Minnesota - Twin Cities REU 2018

1 August 2018

Gao, Li, Mattoo VCIs in P1 × P1 1 / 25

slide-2
SLIDE 2

Outline Preliminaries Determination of VCIs

1 Preliminaries

Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

2 Determination of VCIs

Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Gao, Li, Mattoo VCIs in P1 × P1 2 / 25

slide-3
SLIDE 3

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Projective Space Pn

Definition A projective space Pn over the field C is the set of

  • ne-dimensional subspaces of the vector space Cn+1.
  • The coordinate ring of Pn is S = C[x0, x1, . . . , xn].
  • Grading: Constants have degree 0. Each xi has degree 1.

Gao, Li, Mattoo VCIs in P1 × P1 3 / 25

slide-4
SLIDE 4

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Projective Space Pn

Definition A projective space Pn over the field C is the set of

  • ne-dimensional subspaces of the vector space Cn+1.
  • The coordinate ring of Pn is S = C[x0, x1, . . . , xn].
  • Grading: Constants have degree 0. Each xi has degree 1.

P1 O

[0 : 1] [1 : 1] [2 : 1] [3 : 1] [4 : 1] Gao, Li, Mattoo VCIs in P1 × P1 3 / 25

slide-5
SLIDE 5

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Projective Space Pn

Definition A projective space Pn over the field C is the set of

  • ne-dimensional subspaces of the vector space Cn+1.
  • The coordinate ring of Pn is S = C[x0, x1, . . . , xn].
  • Grading: Constants have degree 0. Each xi has degree 1.

Definition A projective variety X ⊂ Pn is the zero locus of a collection of homogeneous polynomials fα ∈ C[x0, x1, . . . , xn].

Gao, Li, Mattoo VCIs in P1 × P1 3 / 25

slide-6
SLIDE 6

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Biprojective Space P1 × P1

Definition The biprojective space P1 × P1 is the set of equivalence classes: P1 × P1 := {((a0, a1), (b0, b1)) ∈ C2 × C2

  • (a0,a1)=(0,0)

and (b0,b1)=(0,0)}/∼

x ∼ y ⇐ ⇒ x = λy, where x, y ∈ P1, λ ∈ C∗

  • Varieties ↔ zero locus of bihomogenous f ∈ C[x0, x1, y0, y1]
  • Multigrading: deg(xi) = (1, 0), deg(yi) = (0, 1)
  • ex. x2

0y0 + x1x2y1 has degree (2, 1).

Gao, Li, Mattoo VCIs in P1 × P1 4 / 25

slide-7
SLIDE 7

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Biprojective Space P1 × P1

Definition The biprojective space P1 × P1 is the set of equivalence classes: P1 × P1 := {((a0, a1), (b0, b1)) ∈ C2 × C2

  • (a0,a1)=(0,0)

and (b0,b1)=(0,0)}/∼

x ∼ y ⇐ ⇒ x = λy, where x, y ∈ P1, λ ∈ C∗

  • Varieties ↔ zero locus of bihomogenous f ∈ C[x0, x1, y0, y1]
  • Multigrading: deg(xi) = (1, 0), deg(yi) = (0, 1)
  • ex. x2

0y0 + x1x2y1 has degree (2, 1).

  • Irrelevant ideal: B = x0, x1 ∩ y0, y1 ↔ V (B) = ∅
  • Saturation: I : B∞ = {s ∈ S|sBn ⊂ I for some n}

Gao, Li, Mattoo VCIs in P1 × P1 4 / 25

slide-8
SLIDE 8

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

The Nullstellensatz

The Nullstellensatz establishes a correspondence between ideals and varieties: Theorem Let X be a non-empty variety with the coordinate ring S and irrelevant ideal B. If I ⊆ S is a homogeneous ideal, then there is an inclusion-reversing bijective correspondence: {V (I) = ∅}

I

− → ← −

V {radical homogeneous B-saturated ideals ⊂ S}

  • V (I) := zero locus of all f ∈ I
  • I(V (I)) =

√ I

Gao, Li, Mattoo VCIs in P1 × P1 5 / 25

slide-9
SLIDE 9

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Varieties in P1 × P1

Definition P1 × P1 := {((a0, a1), (b0, b1)) ∈ C2 × C2 (a0,a1)=(0,0)

(b0,b1)=(0,0)}/∼ [1 : 0] [1 : 0] [0 : 1] [0 : 1] [1 : 1] [1 : 1] [1 : 2] [1 : 2] [1 : 3] [1 : 3]

Gao, Li, Mattoo VCIs in P1 × P1 6 / 25

slide-10
SLIDE 10

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Varieties in P1 × P1

Definition P1 × P1 := {((a0, a1), (b0, b1)) ∈ C2 × C2 (a0,a1)=(0,0)

(b0,b1)=(0,0)}/∼ [1 : 0] [1 : 0] [0 : 1] [0 : 1] [1 : 1] [1 : 1] [1 : 2] [1 : 2] [1 : 3] [1 : 3] X = ([0 : 1], [0 : 1]) I = x0, y0

Gao, Li, Mattoo VCIs in P1 × P1 6 / 25

slide-11
SLIDE 11

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Varieties in P1 × P1

Definition P1 × P1 := {((a0, a1), (b0, b1)) ∈ C2 × C2 (a0,a1)=(0,0)

(b0,b1)=(0,0)}/∼ [1 : 0] [1 : 0] [0 : 1] [0 : 1] [1 : 1] [1 : 1] [1 : 2] [1 : 2] [1 : 3] [1 : 3] X = ([0 : 1], [0 : 1]) I = x0, y0 ∪([1 : 1], [1 : 1]) ∩x0 − x1, y0 − y1

Gao, Li, Mattoo VCIs in P1 × P1 6 / 25

slide-12
SLIDE 12

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Free Resolution

Definition A free resolution of a module M is an exact sequence of homomorphisms: 0 ← − M

ϕ0

← − F0

ϕ1

← − F1

ϕ2

← − F2 ← − · · · ,

  • im ϕi+1 = ker ϕi at each step
  • every Fi ∼

= Rri is a free module

Gao, Li, Mattoo VCIs in P1 × P1 7 / 25

slide-13
SLIDE 13

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Minimal Free Resolution

Definition A free resolution is minimal if for every ℓ ≥ 1, the nonzero entries of the graded matrix of ϕℓ have positive degree.

  • For each ℓ > 0, ϕℓ takes the standard basis of Fℓ to a

minimal generating set of im(ϕℓ).

  • Unique up to isomorphism.
  • Depends on geometry of points (configuration/cross ratios)

Gao, Li, Mattoo VCIs in P1 × P1 8 / 25

slide-14
SLIDE 14

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Virtual Resolution

Definition A virtual resolution for an ideal I in the biprojective space P1 × P1 is a free complex: 0 ← − I

ϕ0

← − S

ϕ1

← − F1

ϕ2

← − F2

ϕ3

← − · · · such that

  • Fi are free modules for i ≥ 0
  • ann

ker(ϕi)

im(ϕi+1)

  • ⊇ Bl
  • im(ϕ1) : B∞ = I : B∞.

Gao, Li, Mattoo VCIs in P1 × P1 9 / 25

slide-15
SLIDE 15

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Complete and Virtual Complete Intersection

  • X is a complete intersection if I(X) has 2 generators.

V (x0x1) V (y0y1) X =     ([0 : 1], [1 : 0]), ([1 : 0], [1 : 0]), ([0 : 1], [0 : 1]), ([1 : 0], [0 : 1])     = ⇒ I(X) = x0x1, y0y1

Gao, Li, Mattoo VCIs in P1 × P1 10 / 25

slide-16
SLIDE 16

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Complete and Virtual Complete Intersection

  • X is a complete intersection if I(X) has 2 generators.

V (x0x1) V (y0y1) X =     ([0 : 1], [1 : 0]), ([1 : 0], [1 : 0]), ([0 : 1], [0 : 1]), ([1 : 0], [0 : 1])     = ⇒ I(X) = x0x1, y0y1

  • Complete intersection ⇐

⇒ min. free resolution is Koszul: S1 ← S2 ← S1 ← 0

Gao, Li, Mattoo VCIs in P1 × P1 10 / 25

slide-17
SLIDE 17

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Complete and Virtual Complete Intersection

  • X is a complete intersection if I(X) has 2 generators.

V (x0x1) V (y0y1) X =     ([0 : 1], [1 : 0]), ([1 : 0], [1 : 0]), ([0 : 1], [0 : 1]), ([1 : 0], [0 : 1])     = ⇒ I(X) = x0x1, y0y1

  • Complete intersection ⇐

⇒ min. free resolution is Koszul: S1 ← S2 ← S1 ← 0 Definition An ideal I of points in P1 × P1 is a virtual complete intersection (VCI) if I has a short virtual resolution that is Koszul. In particular, V (I) = V (f) ∩ V (g).

Gao, Li, Mattoo VCIs in P1 × P1 10 / 25

slide-18
SLIDE 18

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

VCI Examples

S1 ← S2 ← S1 ← 0 = ⇒ Complete intersection S1 ← S6 ← S8 ← S3 ← 0 = ⇒ Not complete intersection

Gao, Li, Mattoo VCIs in P1 × P1 11 / 25

slide-19
SLIDE 19

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

VCI Examples

S1 ← S2 ← S1 ← 0 = ⇒ Complete intersection S1 ← S2 ← S1 ← 0 S1 ← S6 ← S8 ← S3 ← 0 = ⇒ Not complete intersection S1 ← S2 ← S1 ← 0 = ⇒ Both are VCIs.

Gao, Li, Mattoo VCIs in P1 × P1 11 / 25

slide-20
SLIDE 20

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Generalized B´ ezout’s Theorem

Theorem Let f, g ∈ k[x0, x1, y0, y1] be bihomogeneous forms. If f and g have multidegree (a, b) and (c, d), then |V (f) ∩ V (g)| = ad + bc counting multiplicities.

Gao, Li, Mattoo VCIs in P1 × P1 12 / 25

slide-21
SLIDE 21

Outline Preliminaries Determination of VCIs Projective Space and Varieties Free and Virtual Resolutions Virtual Complete Intersections (VCIs)

Generalized B´ ezout’s Theorem

Theorem Let f, g ∈ k[x0, x1, y0, y1] be bihomogeneous forms. If f and g have multidegree (a, b) and (c, d), then |V (f) ∩ V (g)| = ad + bc counting multiplicities. Red: x0y1 + x1y0: (1, 1) Blue: x0y1 − x1y0: (1, 1) 1 · 1 + 1 · 1 = 2 points. Red: x0x1(y0 − y1): (2, 1) Blue:(x0 − x1)y0y1: (1, 2) 1 · 1 + 2 · 2 = 5 points.

Gao, Li, Mattoo VCIs in P1 × P1 12 / 25

slide-22
SLIDE 22

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Our Main Results

Let X be a set of points in P1 × P1. This is a VCI: each vertical ruling has 2 points.

  • Existence Case: Same

number of points on each ruling.

  • Non existence case: Bound
  • n |X| and maximal

rulings form cross.

  • Further conditions on

VCIs.

Gao, Li, Mattoo VCIs in P1 × P1 13 / 25

slide-23
SLIDE 23

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Our Main Results

Let X be a set of points in P1 × P1.

A (4, 2, 1, 1)-Ferrers Diagram

|X| = 8. We expect 16 points to have a VCI.

  • Existence Case: Same

number of points on each ruling.

  • Non existence case: Bound
  • n |X| and maximal

rulings form cross.

  • Further conditions on

VCIs.

Gao, Li, Mattoo VCIs in P1 × P1 13 / 25

slide-24
SLIDE 24

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Same Cardinality of Rulings

Theorem If X has the same number (n) of points in each vertical (or each horizontal) ruling, it is a VCI.

  • k vertical rulings each having n points

= ⇒ deg(f) = (n, ≤ n), deg(g) = (0, k).

  • Idea: Use Lagrangian interpolation

Gao, Li, Mattoo VCIs in P1 × P1 14 / 25

slide-25
SLIDE 25

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Same Cardinality of Rulings

Theorem If X has the same number (n) of points in each vertical (or each horizontal) ruling, it is a VCI.

  • k vertical rulings each having n points

= ⇒ deg(f) = (n, ≤ n), deg(g) = (0, k).

  • Idea: Use Lagrangian interpolation

Gao, Li, Mattoo VCIs in P1 × P1 14 / 25

slide-26
SLIDE 26

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Same Cardinality of Rulings

Theorem If X has the same number (n) of points in each vertical (or each horizontal) ruling, it is a VCI.

  • k vertical rulings each having n points

= ⇒ deg(f) = (n, ≤ n), deg(g) = (0, k).

  • Idea: Use Lagrangian interpolation

Gao, Li, Mattoo VCIs in P1 × P1 14 / 25

slide-27
SLIDE 27

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Degree Bound Lemma

Setup: f: (a, b)-form, g: (c, d)-form. Assume X = V (f) ∩ V (g). ≤ m points collinear horizontally, ≤ n vertically

Lemma max(a, c) ≥ m and max(b, d) ≥ n. When |X| < mn, we must have a ≥ m, b ≥ n (or c ≥ m, d ≥ n).

n = 3 m = 4

Gao, Li, Mattoo VCIs in P1 × P1 15 / 25

slide-28
SLIDE 28

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Degree Bound Lemma

Setup: f: (a, b)-form, g: (c, d)-form. Assume X = V (f) ∩ V (g). ≤ m points collinear horizontally, ≤ n vertically

Lemma max(a, c) ≥ m and max(b, d) ≥ n. When |X| < mn, we must have a ≥ m, b ≥ n (or c ≥ m, d ≥ n).

n = 3 m = 4

Two cases: deg(f) = (≥ m, ≥ n) deg(g) = ( ? , ? ) deg(f) = (≥ m, ? ) deg(g) = ( ? , ≥ n)

Gao, Li, Mattoo VCIs in P1 × P1 15 / 25

slide-29
SLIDE 29

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition

Theorem If |X| < mn, and there is at least one point in X that is on a horizontal ruling with m points and a vertical ruling with n points, then X is not a VCI. m = 3 n = 4

Gao, Li, Mattoo VCIs in P1 × P1 16 / 25

slide-30
SLIDE 30

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. m = 3 n = 4

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-31
SLIDE 31

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-32
SLIDE 32

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

  • a(s + q) + b(t + p) = |X|

≤ ms + nt − 1+aq + bp

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-33
SLIDE 33

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

  • a(s + q) + b(t + p) = |X|

≤ ms + nt − 1+aq + bp

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-34
SLIDE 34

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

  • a(s + q) + b(t + p) = |X|

≤ ms + nt − 1+aq + bp

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-35
SLIDE 35

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

  • as + bt ≤ ms + nt − 1

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-36
SLIDE 36

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Cross Point Condition: Proof Sketch

Theorem |X| < mn and cross point exists = ⇒ not VCI. s = 1 t = 1 [1 : β] [1 : α]

  • Assume V (f) ∩ V (g) = X. By

B´ ezout, |X| = ad + bc = 7.

  • a ≥ m, b ≥ n.
  • g = (x1 − αx0)(y1 − βy0)g0.
  • Suppose deg(g0) = (p, q).

= ⇒ deg(g) = (t + p, s + q)

  • as + bt ≤ ms + nt − 1

= ⇒ contradiction

Gao, Li, Mattoo VCIs in P1 × P1 17 / 25

slide-37
SLIDE 37

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (a, b)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem Let X be a VCI with |X| < mn.

  • f has degree (m, n) and g has vertical and horizontal

components exactly on rulings with m and n points

  • gcd(m, n) divides |X|
  • If gcd(m, n) = 1: g has degree:

(n−1|X| mod m, m−1|X| mod n)

Gao, Li, Mattoo VCIs in P1 × P1 18 / 25

slide-38
SLIDE 38

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (a, b)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem If |X| < mn: f has degree (m, n) and g has vertical and horizontal components exactly on rulings with m and n points m = 5, n = 4, |X| = 18

Gao, Li, Mattoo VCIs in P1 × P1 19 / 25

slide-39
SLIDE 39

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (a, b)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem If |X| < mn: f has degree (m, n) and g has vertical and horizontal components exactly on rulings with m and n points m = 5, n = 4, |X| = 18 f has degree (5, 4)

Gao, Li, Mattoo VCIs in P1 × P1 19 / 25

slide-40
SLIDE 40

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (a, b)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem If |X| < mn: f has degree (m, n) and g has vertical and horizontal components exactly on rulings with m and n points m = 5, n = 4, |X| = 18 f has degree (5, 4) g has one (1, 0) and one (0, 1) part

Gao, Li, Mattoo VCIs in P1 × P1 19 / 25

slide-41
SLIDE 41

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (m, n)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem If |X| < mn: gcd(m, n) divides |X|

  • By B´

ezout and previous, |X| = md + cn

m = 4,n = 4,|X| = 8 Can be VCI m = 4,n = 4,|X| = 9 Can not be VCI

Gao, Li, Mattoo VCIs in P1 × P1 20 / 25

slide-42
SLIDE 42

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conditions on VCIs

Setup: f: (a, b)-form, g: (c, d)-form. ≤ m points collinear horizontally, ≤ n vertically

Theorem If |X| < mn and gcd(m, n) = 1 g has degree: (n−1|X| mod m, m−1|X| mod n) m = 4, n = 3, |X| = 10 g would have degree (2,1) Impossible, so not VCI

Gao, Li, Mattoo VCIs in P1 × P1 21 / 25

slide-43
SLIDE 43

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Results in Action

8-point VCI 12-point VCI

If |X| < mn, m = 4, n = 4, the only VCI configurations are as shown:

  • By Cross Point Condition, m and

n points share no coordinates

  • By GCD condition, |X| is 8 or 12
  • f has degree (4, 4) and g contains

vertical and horizontal form

  • If |X| = 12 = 4c + 4d, rest of g

must be (1, 0) or (0, 1) form

  • Each such form must have 4

points of X

Gao, Li, Mattoo VCIs in P1 × P1 22 / 25

slide-44
SLIDE 44

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

When values of coordinates matter...

Remark Configuration does not always determine whether a set of points is a VCI. For instance,

In general, not a VCI.

( 1

2 , 2)

( 1

3 , 3)

( 1

4 , 4)

(0, 0) (0, 1), (1, 1)

Red:(2, 1); Blue:(2, 2).

Gao, Li, Mattoo VCIs in P1 × P1 23 / 25

slide-45
SLIDE 45

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

When values of coordinates matter...

Remark Configuration does not always determine whether a set of points is a VCI. For instance,

In general, not a VCI.

( 1

2 , 2)

( 1

3 , 3)

( 1

4 , 4)

(0, 0) (0, 1), (1, 1)

Red:(2, 1); Blue:(2, 2).

Gao, Li, Mattoo VCIs in P1 × P1 23 / 25

slide-46
SLIDE 46

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

When values of coordinates matter...

Remark Configuration does not always determine whether a set of points is a VCI. For instance,

In general, not a VCI.

( 1

2 , 2)

( 1

3 , 3)

( 1

4 , 4)

(0, 0) (0, 1), (1, 1)

Red:(2, 1); Blue:(2, 2).

Gao, Li, Mattoo VCIs in P1 × P1 23 / 25

slide-47
SLIDE 47

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conclusion

  • In P

n, virtual resolutions better encode geometry.

Gao, Li, Mattoo VCIs in P1 × P1 24 / 25

slide-48
SLIDE 48

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conclusion

  • In P

n, virtual resolutions better encode geometry.

  • Exists 1-2-1 virtual resolution ⇐

⇒ VCI

Gao, Li, Mattoo VCIs in P1 × P1 24 / 25

slide-49
SLIDE 49

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conclusion

  • In P

n, virtual resolutions better encode geometry.

  • Exists 1-2-1 virtual resolution ⇐

⇒ VCI

  • Our results:

1 Same # of points on each ruling =

⇒ VCI

2 When |X| < mn, restrictions on what VCIs must look like 3 Actual values of the coordinates can affect VCI, too.

Gao, Li, Mattoo VCIs in P1 × P1 24 / 25

slide-50
SLIDE 50

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Conclusion

  • In P

n, virtual resolutions better encode geometry.

  • Exists 1-2-1 virtual resolution ⇐

⇒ VCI

  • Our results:

1 Same # of points on each ruling =

⇒ VCI

2 When |X| < mn, restrictions on what VCIs must look like 3 Actual values of the coordinates can affect VCI, too.

  • Future work:

1 Continue Classification 2 Methods for finding f and g

Gao, Li, Mattoo VCIs in P1 × P1 24 / 25

slide-51
SLIDE 51

Outline Preliminaries Determination of VCIs Overview VCI Existence Cases VCI Non-Existence Conditions on VCIs Conclusion

Acknowledgements

We would like to thank Christine and Mike for their continual guidance, support, and encouragement. Thank you to the other mentors and TAs for their help in the REU and to the NSF for funding us.

Gao, Li, Mattoo VCIs in P1 × P1 25 / 25