An object in a monoidal category A ∈ 𝒞 is right dualizable if there is an object A* and morphisms εA : A* ⊗ A → 1 ηA : 1 → A ⊗ A* such that ηA ⊗ 1A* ∘ α ∘ 1A* ⊗ εA = λA*−1 ∘ ρA* and 1A ⊗ ηA ∘ α ∘ εA ⊗ 1A = ρA−1 ∘ λA
Invertible: In the case that *ε,
η* are isomorphisms, we call A an invertible
object.
Other people will call this a left dual. I'm using right dual to signify that the unit-like map η : 1 → A ⊗ A* puts the dual on the right of the tensor product, with the intuition that a unit should be something like the existence of the object.
A category 𝒞 is right rigid if every object has a right dual. Similarly it is left rigid if every object has a left dual. A rigid category is one that is both left and right rigid.
Let 𝒞 be a category with a terminal object •. An object A is simple if every coequalizer of the form E ⇉ A → Q has Q ≅ 1 or Q ≅ A
An abelian category 𝒞 is semisimple, if every object is a direct sum of finitely many simple objects.
A fusion category C is a rigid, semisimple, Vect-enriched monoidal category, with only finitely many simple objects (up to isomorphism), and such that the endomorphisms of the unit object just form the ground field k.
The goal of a fusion category is to study how objects combine
Since every object can be written as a direct sum of finitely many simple objects, and there are only finitely many simple objects, Xi, we have a kind of structure coefficient equivalence that we can write:
Xi ⊗ Xj ≅ ⨁kNijkXk. where Nijk ∈ ℕ ≥ 0.
Endomorphisms of unit:
Tannaka duality: Every fusion category
is the representation category of a weak Hopf algebra.
A ribbon category is a rigid braided monoidal category (ℬ, ⊗, σ) along with a natural isomorphism θ : I → I from the identity functor to itself, called a twist such that
θA ⊗ B = σBA ∘ σAB ∘ θA ⊗ θB θ1 = id, where 1 is the monoidal unit θA* = θA*
These categories look at how things knot and twist together
Twist: The twist operation is meant to
account for the spin of a particle. As it turns out,
the world lines of anyons shouldn't really be thought of as world
lines, but rather world ribbons, where the extra
dimension corresponds to the spin
Quantum groups: Finite-dimensional
representations of a quantum group form ribbon categories.
A modular tensor category is a ribbon and fusion category 𝒞 which satisfies a non-degeneracy condition
These are similar to ribbon categories, but prevent certain kinds of singularities that can show up in the knots
Let {Xi} be a set of representatives of the simple objects.
Trace: Let f : A → A be an
endomorphism. We define the trace of f by
Tr(f) = ε ∘ (idA* ⊗ (f ∘ θA)) ∘ (idA ⊗ θA*) ∘ σA*A ∘ η
It's similar to the ordinary trace, but we have a twisting action on both strands. Note that Tr(f) : 1 → 1 is a member of the endomorphism ring of the vacuum, which by the definition of a fusion category, is isomorphic to the ground field. So Tr(f) ∈ k.
In the case that your category is the standard category of vector spaces, you get back the ordinary trace (because the twist is trivial there).
S-matrix: To form the Smatrix, we take
the trace of a double braiding:
Sij = Tr(σAjAi ∘ σAiAj)
Non-degeneracy: The non-degeneracy
condition is equivalent to the Smatrix being non-degenerate.
TQFTs: Modular tensor categories
describe 2 + 1 dimensional topological
quantum field theory.
Modular group: