The combinatorial Maslov index of a domain #
Lipshitz's index formula computes the expected dimension of the moduli space of holomorphic
disks in a Whitney class φ from its domain D alone:
μ(φ) = e(D) + n_x(D) + n_y(D). This file defines the right-hand side for the abstract region
incidence data TauCeti.HeegaardRegionSystem and proves that it is additive under juxtaposition
of domains, a property needed for a relative grading once independence of the choice of domain is
established.
Every intersection point p is a corner of four regions (with repetition): the regions on the
two sides of the α-arc ending at p and of the α-arc starting at p. The point measure
n_p(D) is the average of the multiplicities of D at these four corners, and for a generator
x = {x₁, …, xₙ} one puts n_x(D) = ∑ n_{xᵢ}(D). With the standard combinatorial convention
that every corner contributes one quarter, the Euler measure of a region R with k corners
is χ(R) - k / 4; it is extended linearly to domains. The Euler characteristics χ(R) of the
regions are not determined by the incidence data, so they are an explicit argument.
The key combinatorial fact is that for a domain D from x to y and a domain E from y to
w, n_x(E) + n_w(D) = n_y(D) + n_y(E). Both sides differ by a corner-averaged intersection
number of ∂D ∩ α with ∂E ∩ β, computed once along the α-curves and once along the
β-curves, and the two computations agree up to sign at every crossing.
Main definitions #
TauCeti.HeegaardRegionSystem.cornerRegion: the four regions with a corner at a point.TauCeti.HeegaardRegionSystem.pointMeasure: the point measuren_p(D).TauCeti.HeegaardRegionSystem.generatorPointMeasure: the point measuren_x(D)of a generator.TauCeti.HeegaardRegionSystem.cornerCount: the number of corners of a region.TauCeti.HeegaardRegionSystem.eulerMeasure: the Euler measuree(D).TauCeti.HeegaardRegionSystem.maslovIndex: the combinatorial Maslov indexe(D) + n_x(D) + n_y(D).
Main results #
TauCeti.HeegaardRegionSystem.pointMeasure_apply_beta: the point measure is also the average over the corners on the two sides of theβ-arcs.TauCeti.HeegaardRegionSystem.eulerMeasure_eq_sub_sum_pointMeasure: the Euler measure is∑ χ(R) D(R) - ∑ₚ n_p(D).TauCeti.HeegaardRegionSystem.IsDomainBetween.generatorPointMeasure_add_generatorPointMeasure:n_x(E) + n_w(D) = n_y(D) + n_y(E)forDfromxtoyandEfromytow.TauCeti.HeegaardRegionSystem.IsDomainBetween.maslovIndex_add: the Maslov index is additive under juxtaposition.TauCeti.HeegaardRegionSystem.IsDomainBetween.maslovIndex_self_eq: a domain whose boundary is a sum of whole curves has the same Maslov index at any two generators joined by a domain.
References #
- R. Lipshitz, A cylindrical reformulation of Heegaard Floer homology, Geom. Topol. 10 (2006), arXiv:math/0502404, §4, where the index formula and its additivity are proved; see also the erratum arXiv:1301.4919.
- S. Sarkar, Maslov index formulas for Whitney n-gons, J. Topol. 4 (2011), arXiv:math/0609673, for the combinatorial proof of the additivity of point measures.
The four regions with a corner at the intersection point p, with repetition: the regions
to the left and to the right of the α-arc ending at p, then those of the α-arc starting at
p.
Equations
- H.cornerRegion p = ![H.alphaLeft ((Equiv.symm H.alphaNext) p), H.alphaRight ((Equiv.symm H.alphaNext) p), H.alphaLeft p, H.alphaRight p]
Instances For
The point measure n_p(D): the average multiplicity of the domain D at the four corners
at the intersection point p.
Equations
- H.pointMeasure = { toFun := fun (D : Region → ℤ) (p : Point) => (∑ i : Fin 4, ↑(D (H.cornerRegion p i))) / 4, map_zero' := ⋯, map_add' := ⋯ }
Instances For
The four corners at a crossing are also the regions on the two sides of the β-arc ending
there and of the β-arc starting there, so the point measure can be read off along β.
The point measure n_x(D) = ∑ᵢ n_{xᵢ}(D) of a domain D at a generator x.
Equations
- H.generatorPointMeasure x = { toFun := fun (D : Region → ℤ) => ∑ i : Fin n, H.pointMeasure D (H.point x i), map_zero' := ⋯, map_add' := ⋯ }
Instances For
The point measure of a generator is the pairing of its 0-chain with the point
measures.
The point measures of two juxtaposable domains satisfy n_x(E) + n_w(D) = n_y(D) + n_y(E)
when D connects x to y and E connects y to w.
A domain whose boundary is a sum of whole curves has the same point measure at any two generators joined by a domain.
The number of corners of the region r, counted with multiplicity over the four corners at
every intersection point.
Equations
- H.cornerCount r = {c : Point × Fin 4 | H.cornerRegion c.1 c.2 = r}.card
Instances For
The Euler measure e(D) = ∑ D(R) (χ(R) - k(R) / 4) of a domain, where χ(R) is the Euler
characteristic of the region R, supplied as χ, and k(R) is its number of corners.
Equations
- H.eulerMeasure χ = { toFun := fun (D : Region → ℤ) => ∑ r : Region, ↑(D r) * (↑(χ r) - ↑(H.cornerCount r) / 4), map_zero' := ⋯, map_add' := ⋯ }
Instances For
The Euler measure of a single region R is χ(R) - k(R) / 4.
Each corner contributes a quarter of the multiplicity at its region to the point measure at
its vertex, so the Euler measure is ∑ χ(R) D(R) - ∑ₚ n_p(D).
The combinatorial Maslov index μ(D) = e(D) + n_x(D) + n_y(D) of a domain D from the
generator x to the generator y, for the Euler characteristics χ of the regions.
Equations
- H.maslovIndex χ x y = H.eulerMeasure χ + H.generatorPointMeasure x + H.generatorPointMeasure y
Instances For
Reversing a domain negates its Maslov index.
The Maslov index is additive under juxtaposition of domains: if D connects x to y and
E connects y to w, then μ(D + E) = μ(D) + μ(E).
A domain whose boundary is a sum of whole curves has the same Maslov index at any two generators joined by a domain.