Dual hypergroup of a commutative Hecke algebra
DefinitionQSDMixing_GU_Supplement_HeckeDualDatafourier-analysismarkov-chainp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-mixing-timequantum-state-discrimination
For a commutative Hecke algebra , dual-hypergroup data consist of its primitive orthogonal idempotents , their positive multiplicities , and Krein parameters satisfying
The dual convolution is , and the Haar law is . The definition records the hypergroup probability, associativity, identity, and stationarity laws. It also defines , convolution powers, and the corresponding total-variation mixing time.
This is the representation-theoretic object used to turn the Schur-channel walk into the classical walk of Theorem 4.
Definition code
import Mathlib
import Definitions.Def_QSDMixing_GU_hecke
import Definitions.Def_QSDMixing_GU_Supplement_FiniteProbability
set_option autoImplicit false
open scoped MatrixOrder ComplexOrder
namespace QSDMixing.GU.Supplement
noncomputable local instance heckeDualDecidableEq {α : Type} : DecidableEq α :=
Classical.decEq α
/-- Primitive-idempotent and Krein-parameter data for the dual hypergroup of a commutative
Hecke algebra. The fields spell out the normalization used in §3.2, equations (80)--(85).
Zhou--Leditzky, *A mixing time method for estimating the sample complexity of quantum state
discrimination*, arXiv:2609.40182 (held version, sha256 62fb68dd…), §3.2, pp.19--20. -/
structure HeckeDualData {G : Type} [Group G] [Finite G] (H : Subgroup G)
(Ω : Type) [Fintype Ω] where
E : Ω → Matrix (G ⧸ H) (G ⧸ H) ℂ
m : Ω → ℝ
q : Ω → Ω → Ω → ℝ
identity : Ω
E_mem : ∀ i, E i ∈ QSDMixing.GU.hecke H
E_pos : ∀ i, (E i).PosSemidef
E_mul : ∀ i j, E i * E j = if i = j then E i else 0
E_sum : ∑ i, E i = 1
primitive : ∀ i (P : Matrix (G ⧸ H) (G ⧸ H) ℂ),
P ∈ QSDMixing.GU.hecke H → P * P = P → P * E i = P → P = 0 ∨ P = E i
m_pos : ∀ i, 0 < m i
m_trace : ∀ i, Matrix.trace (E i) = (m i : ℂ)
schur_formula : ∀ i j,
(fun x y => E i x y * E j x y) =
(Fintype.card (G ⧸ H) : ℂ)⁻¹ •
∑ k, ((q i j k : ℝ) : ℂ) • E k
conv : Ω → Ω → Ω → ℝ
conv_formula : ∀ i j k,
conv i j k = m k * q i j k / (m i * m j)
conv_nonneg : ∀ i j k, 0 ≤ conv i j k
conv_sum_one : ∀ i j, ∑ k, conv i j k = 1
conv_assoc : ∀ i j k l,
∑ a, conv i j a * conv a k l = ∑ a, conv j k a * conv i a l
conv_left_id : ∀ i j, conv identity i j = if i = j then 1 else 0
conv_right_id : ∀ i j, conv i identity j = if i = j then 1 else 0
haar : FiniteProbability Ω
haar_formula : ∀ i, haar i = m i / (Fintype.card (G ⧸ H) : ℝ)
haar_stationary : ∀ i k, ∑ j, haar j * conv i j k = haar k
/-- The primitive-idempotent map `Ξ(q)=∑ᵢ(qᵢ/mᵢ)Eᵢ`. -/
noncomputable def HeckeDualData.xi {G : Type} [Group G] [Finite G] {H : Subgroup G}
{Ω : Type} [Fintype Ω] (D : HeckeDualData H Ω) (p : Ω → ℝ) :
Matrix (G ⧸ H) (G ⧸ H) ℂ :=
∑ i, (((p i / D.m i : ℝ) : ℂ)) • D.E i
/-- Convolution of probability laws on the dual hypergroup. -/
noncomputable def HeckeDualData.measureConvolution {G : Type} [Group G] [Finite G]
{H : Subgroup G} {Ω : Type} [Fintype Ω] (D : HeckeDualData H Ω)
(p r : Ω → ℝ) (k : Ω) : ℝ :=
∑ i, ∑ j, p i * r j * D.conv i j k
/-- The `t`-fold dual-hypergroup convolution power. -/
noncomputable def HeckeDualData.convolutionPower {G : Type} [Group G] [Finite G]
{H : Subgroup G} {Ω : Type} [Fintype Ω] (D : HeckeDualData H Ω)
(p : Ω → ℝ) : ℕ → Ω → ℝ
| 0 => fun z => if z = D.identity then 1 else 0
| t + 1 => D.measureConvolution (D.convolutionPower p t) p
/-- The total-variation mixing time of the dual-hypergroup walk, starting at its identity. -/
noncomputable def HeckeDualData.mixingTime {G : Type} [Group G] [Finite G]
{H : Subgroup G} {Ω : Type} [Fintype Ω] (D : HeckeDualData H Ω)
(p : Ω → ℝ) (ε : ℝ) : ℕ∞ :=
sInf {T : ℕ∞ | ∃ t : ℕ, 1 ≤ t ∧ T = t ∧
totalVariation (D.convolutionPower p t) D.haar ≤ ε}
end QSDMixing.GU.Supplement
Source
Zhou, Leditzky, A mixing time method for estimating the sample complexity of quantum state discrimination, arXiv:2609.40182 (held version, sha256 62fb68dd…), §3.2, pp.19–20, eqs. (80)–(85)