Carleman flux criterion for a general hop: boundary layers and the recursion
DefinitionChapterCarlemanGeneralHopPromotedFormal definitions for the timepiece Lean 4 formalization (source chapter BookProof/ChapterCarlemanGeneralHop.lean): the boundary layers of a cube for a general hop α ↦ α + p h − m h, built on the re-carried Hermite-Carleman and two-step definitions (the served Def_ChapterHermiteCarlemanEsa copy predates most of that vocabulary, and published bundles are immutable, so the definitions reach the platform under new bundle names). This bundle carries the chapter's definitions: the hop hshift, the shift family hopB, the outgoing and incoming boundary layers obd/ibd of a cube, the single-hop contributions ltG/hopG, and the general-hop recursion LadderRecH. The chapter's theorem facts stay problem nodes. Consumers add a single import Definitions.Def_ChapterCarlemanGeneralHopPromoted.
-- §2.13 new-name route (Session 40): the platform's served `Def_ChapterCarlemanGeneralHop` copy is a
-- skeleton (no declarations at all), so `open BookProof.CarlemanGeneralHop` in every consumer statement
-- dies with `unknown namespace`. `submit-definition` cannot update an existing
-- name, so this bundle carries the chapter's DEFINITIONS in the original
-- namespace; its theorem facts stay problem nodes. Consumers add a single
-- `import Definitions.Def_ChapterCarlemanGeneralHopPromoted`.
-- Left to the problem nodes on purpose (theorem facts):
-- hshift_apply
-- hshift_hshift
-- hshift_le
-- mem_hopB
-- ltG_eq_conj_rtG
-- sum_ltG
-- mem_obd
-- mem_ibd
-- amp_eq_zero_of_not_mem_obd
-- in_mem_ibd
-- flux_bound_gen
-- obd_image_multiplicity
-- ibd_multiplicity
-- obd_mass_le
import Definitions.Def_ChapterCarlemanTwoStepPromoted
import Definitions.Def_ChapterHermiteCarlemanEsaPromoted
import Mathlib
/-!
# A Carleman criterion for general lattice hops
`BookProof.ChapterHermiteCarlemanEsa` and `BookProof.ChapterCarlemanTwoStep` prove
Carleman (flux) criteria for recursions whose hops move a **single** excitation number,
by one or by two. That covers every **mode-diagonal** quadratic Hamiltonian. A quadratic
Hamiltonian which couples two *distinct* modes — `xᵢxⱼ`, `πᵢπⱼ`, `xᵢπⱼ` with `i ≠ j` —
produces hops `α ↦ α ± (eᵢ + eⱼ)` and `α ↦ α ± (eᵢ − eⱼ)`, and the second kind is **not**
monotone: the shift lowers one coordinate while raising another.
This module runs the flux argument for a hop of the completely general shape
`α ↦ α + p − m` (`hshift`), with `p` and `m` multi-indices.
## What is proved
* `hshift`, `hshift_hshift` — the shift and its inverse on the set where it is defined.
* `rtG`, `ltG`, `hopB`, `sum_ltG`, `sum_hop_im` — **the abstract flux cancellation.** For
a Hermitian hop family, the contributions from pairs `α, α + p − m` which both lie in a
finite set `A` cancel in the imaginary part, so the imaginary part of the total is
carried by two boundary layers: the *outgoing* layer `A \ B` (points of `A` whose image
leaves `A`) and the *incoming* layer `B \ A` (points outside `A` whose image lands in
`A`). For a monotone hop (`m = 0`) the incoming layer is empty and this specialises to
the situation of the two earlier modules.
-/
namespace BookProof.CarlemanGeneralHop
open Finset
open BookProof.HermiteCarleman
open BookProof.CarlemanTwoStep
noncomputable section
variable {d : ℕ}
/-! ## 1. The general shift -/
/-- The lattice hop `α ↦ α + p − m` (truncated subtraction; it is used only where
`m ≤ α`). -/
def hshift (p m a : Fin d →₀ ℕ) : Fin d →₀ ℕ := a + p - m
/-! ## 2. The abstract flux cancellation -/
variable {u : (Fin d →₀ ℕ) → ℂ}
/-- The raising contribution of the hop `α ↦ α + p − m`. -/
def rtG (u : (Fin d →₀ ℕ) → ℂ) (w : ℂ) (c : (Fin d →₀ ℕ) → ℝ) (p m a : Fin d →₀ ℕ) : ℂ :=
(starRingEnd ℂ) w * ((c a : ℝ) : ℂ) * (starRingEnd ℂ) (u a) * u (hshift p m a)
/-- The lowering contribution, i.e. the contribution of the opposite hop
`α ↦ α − p + m`. -/
def ltG (u : (Fin d →₀ ℕ) → ℂ) (w : ℂ) (c' : (Fin d →₀ ℕ) → ℝ) (p m a : Fin d →₀ ℕ) : ℂ :=
w * ((c' a : ℝ) : ℂ) * (starRingEnd ℂ) (u a) * u (hshift m p a)
/-- The set of points that hop **into** `A`. -/
def hopB (A : Finset (Fin d →₀ ℕ)) (p m : Fin d →₀ ℕ) : Finset (Fin d →₀ ℕ) :=
(A.filter (fun a => ∀ k, p k ≤ a k)).image (hshift m p)
/-! ## 3. The boundary layers of a cube -/
/-- The **outgoing** boundary layer of the cube for the hop `α ↦ α + p − m`: the points of
the cube at which the hop is defined and some coordinate is within `p` of the top. -/
def obd (d N : ℕ) (p m : Fin d →₀ ℕ) : Finset (Fin d →₀ ℕ) :=
(cube d N).filter (fun a => (∀ k, m k ≤ a k) ∧ ∃ k, N < a k + p k)
/-- The **incoming** boundary layer: points just outside the cube whose hop image can land
inside it. It is empty for a monotone hop (`m = 0`). -/
def ibd (d N : ℕ) (m : Fin d →₀ ℕ) : Finset (Fin d →₀ ℕ) :=
(cube d (N + 1)).filter (fun b => (∀ k, m k ≤ b k) ∧ ¬ (∀ k, b k ≤ N))
/-! ## 4. The flux bound -/
/-! ## 5. Bessel's inequality for the two boundary layers -/
/-! ## 6. The recursion and the criterion -/
/-- **A general-hop ladder recursion.** A real diagonal `lam`, and a finite family of
Hermitian hops `α ↦ α + p h − m h` with constant amplitudes `w h`. -/
def LadderRecH (u : (Fin d →₀ ℕ) → ℂ) (lam : (Fin d →₀ ℕ) → ℝ) {ι : Type*} [Fintype ι]
(p m : ι → (Fin d →₀ ℕ)) (c c' : ι → (Fin d →₀ ℕ) → ℝ) (w : ι → ℂ) (z : ℂ) : Prop :=
∀ a : Fin d →₀ ℕ,
((lam a : ℝ) : ℂ) * u a
+ ∑ h : ι, ((starRingEnd ℂ) (w h) * ((c h a : ℝ) : ℂ) * u (hshift (p h) (m h) a)
+ w h * ((c' h a : ℝ) : ℂ) * u (hshift (m h) (p h) a))
= z * u a
variable {ι : Type*} [Fintype ι] {lam : (Fin d →₀ ℕ) → ℝ} {p m : ι → (Fin d →₀ ℕ)}
{c c' : ι → (Fin d →₀ ℕ) → ℝ} {w : ι → ℂ} {z : ℂ}
end
end BookProof.CarlemanGeneralHop