ConceptComplete

Homological Mirror Symmetry - Examples and Constructions

Proving homological mirror symmetry requires sophisticated techniques from symplectic geometry, algebraic geometry, and category theory. We examine proven cases and construction methods that have established HMS in various contexts.

ExampleHMS for Elliptic Curves (Polishchuk-Zaslow)

For an elliptic curve E=C/ΛE = \mathbb{C}/\Lambda with dual E^=C/Λ\hat{E} = \mathbb{C}/\Lambda^\vee:

B-side: Db(Coh(E))D^b(Coh(E)) has generators:

  • OE\mathcal{O}_E: structure sheaf
  • Op\mathcal{O}_p: skyscraper at point pp

A-side: DFuk(E^)D\mathcal{F}uk(\hat{E}) has generators:

  • Lagrangian tori (1-cycles)
  • Lagrangian circles

The equivalence sends line bundles to Lagrangian tori via the Abel-Jacobi map, and point sheaves to specific Lagrangian cycles. This was the first rigorous proof of HMS.

Quartic Surfaces and HMS

Seidel proved HMS for quartic K3 surfaces using symplectic techniques.

ExampleQuartic K3 Surfaces

A quartic K3 surface SP3S \subset \mathbb{P}^3 defined by a degree 4 polynomial satisfies: Db(Coh(S))DFuk(S)D^b(Coh(S)) \simeq D\mathcal{F}uk(S^\vee)

where SS^\vee is the mirror K3 (another quartic with swapped lattice polarization).

Key technique: Use symplectic Khovanov homology and categorical localization to construct the equivalence explicitly.

The proof involves constructing a sequence of auto-equivalences (spherical twists) that transform the standard generators into the desired mirror generators.

Toric Varieties

HMS for toric varieties provides a framework where both sides can be computed explicitly.

DefinitionToric HMS

For a toric Fano variety XΣX_\Sigma defined by a fan Σ\Sigma: Db(Coh(XΣ))Fuk(MΣ)D^b(Coh(X_\Sigma)) \simeq \mathcal{F}uk(M_\Sigma)

where MΣM_\Sigma is the Landau-Ginzburg mirror, a space with superpotential W:(C)nCW: (\mathbb{C}^*)^n \to \mathbb{C} encoding toric data.

Objects on the A-side are thimbles: Lagrangian submanifolds defined as gradient flow lines of Re(W)\text{Re}(W).

Homological Mirror Symmetry for Weighted Projective Spaces

Weighted projective spaces provide computable examples beyond the classical cases.

Example$\mathbb{P}^{1,1,2}$ Weighted Projective Line

The weighted projective line P1,1,2\mathbb{P}^{1,1,2} is mirror to a Landau-Ginzburg model ( mathbbC,W)(\ mathbb{C}^*, W) with superpotential: W(x)=x+1x2W(x) = x + \frac{1}{x^2}

The derived category Db(Coh(P1,1,2))D^b(Coh(\mathbb{P}^{1,1,2})) (generated by line bundles O,O(1),O(2)\mathcal{O}, \mathcal{O}(1), \mathcal{O}(2)) is equivalent to the Fukaya-Seidel category of (mathbbC,W)(mathbb{C}^*, W).

Critical points of WW correspond to exceptional collections of sheaves.

SYZ Construction and HMS

The Strominger-Yau-Zaslow conjecture provides a geometric approach to HMS via special Lagrangian fibrations.

Remark

If XX admits a special Lagrangian fibration π:XB\pi: X \to B with fibers TnT^n, the mirror XX^\vee is constructed as the moduli space of pairs (L,)(L, \nabla) where:

  • LL is a fiber of π\pi
  • \nabla is a flat U(1)U(1)-connection on LL

The HMS equivalence relates:

  • Coherent sheaves on XX^\vee \leftrightarrow Lagrangian sections of π\pi

Derived Equivalences via Fourier-Mukai

Fourier-Mukai theory provides a framework for constructing derived equivalences.

DefinitionFourier-Mukai Kernel

An equivalence Φ:Db(Coh(X))Db(Coh(Y))\Phi: D^b(Coh(X)) \to D^b(Coh(Y)) is Fourier-Mukai if: Φ(F)=RπY(πXFK)\Phi(\mathcal{F}) = R\pi_{Y*}(\pi_X^*\mathcal{F} \otimes \mathcal{K})

where KDb(Coh(X×Y))\mathcal{K} \in D^b(Coh(X \times Y)) is the kernel.

Orlov's theorem states that all equivalences between derived categories of smooth projective varieties are Fourier-Mukai.

For HMS, the Fukaya-Mukai transform should similarly be expressible via a kernel in Db(Coh(X))×DFuk(X)D^b(Coh(X)) \times D\mathcal{F}uk(X^\vee), though making this precise requires substantial technicalities.

Computational Tools

Proving HMS typically requires:

  1. Semiorthogonal decompositions: Breaking categories into simpler pieces
  2. Exceptional collections: Ordered sets of objects generating the category
  3. Window categories: Subcategories described by inequalities
  4. Matrix factorizations: For Landau-Ginzburg models

These tools allow explicit verification of categorical equivalences, turning abstract conjectures into computational theorems.