Carleman simplex: mass estimates, polynomial action and the criterion
DefinitionChapterCarlemanSimplexPromotedFormal definitions for the timepiece Lean 4 formalization (source chapter BookProof/ChapterCarlemanSimplex.lean): the Carleman simplex criterion. The platform's served Def_ChapterCarlemanSimplex copy is a skeleton under a different namespace, so every consumer statement that says open BookProof.CarlemanSimplex dies with unknown namespace, and published bundles are immutable, so the fix is a new bundle name that keeps the original namespace (the new-name route). This bundle carries the chapter's definitions: the total degree deg, the simplex shells simplexF with their interiors sInn and boundaries sBd, the hop contributions rtermP/ltermP, the mode-exchange layer shiftm/rcm/mterm, the pair-hop amplitudes pvec/rcp/lcp, and the general quadratic recursion LadderRecQ. The chapter's theorem facts stay problem nodes. Consumers add a single import Definitions.Def_ChapterCarlemanSimplexPromoted.
-- §2.13 new-name route (Session 40): the platform's served `Def_ChapterCarlemanSimplex` copy is a
-- skeleton (no declarations at all), so `open BookProof.CarlemanSimplex` 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_ChapterCarlemanSimplexPromoted`.
-- Left to the problem nodes on purpose (theorem facts):
-- deg_add
-- apply_le_deg
-- deg_single
-- tsub_add_cancel_of_le'
-- deg_tsub_of_le
-- mem_simplexF
-- mem_sInn
-- mem_sBd
-- sBd_subset_simplexF
-- sum_simplex_split
-- ltermP_shift
-- sum_ltermP
-- rcm_of_zero
-- deg_shiftm
-- shiftm_apply_self
-- rcm_shiftm
-- sBd_mass_le
-- shifted_sBd_mass_le
-- rcp_nonneg
-- deg_pvec
-- flux_bound_on
-- flux_identityQ
-- not_summable_inv_natCast_add_two
import Definitions.Def_ChapterCarlemanTwoStepPromoted
import Definitions.Def_ChapterHermiteCarlemanEsaPromoted
import Mathlib
/-!
# A Carleman criterion on simplex shells: hops which couple distinct modes
`BookProof.ChapterHermiteCarlemanEsa` and `BookProof.ChapterCarlemanTwoStep` run the
Carleman flux argument on **cubes** `{α : ∀ i, αᵢ ≤ N}`, for hops which move a *single*
excitation number: `α ↦ α ± eᵢ` and `α ↦ α ± 2eᵢ`. That is exactly the ladder structure
of a quadratic Hamiltonian which does not couple distinct modes.
A general real quadratic Hamiltonian does couple them: `xᵢxⱼ`, `πᵢπⱼ` and `xᵢπⱼ` with
`i ≠ j` produce the hops `α ↦ α ± (eᵢ + eⱼ)` and `α ↦ α + eᵢ − eⱼ`. On a cube the
bookkeeping of such hops is awkward — a hop leaves a cube through *two* faces at once,
and the mixed hop `α ↦ α + eᵢ − eⱼ` leaves it through one face while entering through
another.
This module reruns the argument on the **simplex shells** `{α : |α| ≤ N}`, where
`|α| = ∑ᵢ αᵢ` is the total degree. The exhaustion is adapted to the grading by total
excitation number, and everything becomes uniform:
* a hop which *raises* the total degree by `k` (`α ↦ α + P` with `|P| = k`) leaks only
through the shell `{N − k < |α| ≤ N}`, which meets at most `k` of the shells;
* a hop which *preserves* the total degree (`α ↦ α + eᵢ − eⱼ`) never leaves a shell, so
it contributes **nothing at all** to the flux: the corresponding sum over a shell is
real as soon as its amplitude matrix is Hermitian.
## What is proved
* `deg`, `simplexF`, `sInn`, `sBd` — the total degree, the simplex shells and their
interiors and boundaries.
* `sum_simplex_hop_im` — **the abstract flux cancellation** for a hop `α ↦ α + P` of an
arbitrary shift `P`: the interior contributions occur in conjugate pairs, so only the
boundary shell contributes to the imaginary part.
* `sum_mterm_im` — **the degree-preserving hops carry no flux**: for a Hermitian
amplitude matrix the total contribution of the hops `α ↦ α + eᵢ − eⱼ` over a shell is
real.
* `LadderRecQ`, `flux_identityQ` — the recursion of a *general* quadratic ladder — one
step, two steps, pair creation/annihilation and mode exchange — and its flux identity.
* `flux_bound_on`, `sBd_multiplicity`, `shifted_sBd_multiplicity` — the flux bound and
the summability of the boundary mass (each index lies in at most `k` boundary shells).
* `ladderQ_eq_zero` — **the criterion.** A square-summable family satisfying the general
quadratic recursion, with a real diagonal, constant amplitudes and a Hermitian exchange
matrix, at a point off the real axis, vanishes. The Carleman divergence used is
`∑ 1/(N+2) = ∞`.
Everything is `sorry`-free and `axiom`-free.
-/
namespace BookProof.CarlemanSimplex
open Finset
open BookProof.HermiteCarleman BookProof.CarlemanTwoStep
noncomputable section
variable {d : ℕ}
/-! ## 1. The simplex shells -/
/-- The total degree `|α| = ∑ᵢ αᵢ` of a multi-index. -/
def deg (a : Fin d →₀ ℕ) : ℕ := ∑ i, a i
/-- The simplex shell `{α : |α| ≤ N}` of multi-indices, as a finite set. -/
def simplexF (d N : ℕ) : Finset (Fin d →₀ ℕ) := (cube d N).filter (fun a => deg a ≤ N)
/-- The part of a shell which can still be raised by `k` degrees without leaving it. -/
def sInn (d N k : ℕ) : Finset (Fin d →₀ ℕ) := (simplexF d N).filter (fun a => deg a + k ≤ N)
/-- The `k`-thick boundary of a shell. -/
def sBd (d N k : ℕ) : Finset (Fin d →₀ ℕ) := (simplexF d N).filter (fun a => N < deg a + k)
/-! ## 2. The abstract flux cancellation for a raising hop -/
variable {u : (Fin d →₀ ℕ) → ℂ}
/-- The raising contribution of the hop `α ↦ α + P`. -/
def rtermP (u : (Fin d →₀ ℕ) → ℂ) (w : ℂ) (rc : (Fin d →₀ ℕ) → ℝ) (P : Fin d →₀ ℕ)
(a : Fin d →₀ ℕ) : ℂ :=
(starRingEnd ℂ) w * ((rc a : ℝ) : ℂ) * (starRingEnd ℂ) (u a) * u (a + P)
/-- The lowering contribution of the hop `α ↦ α − P`. -/
def ltermP (u : (Fin d →₀ ℕ) → ℂ) (w : ℂ) (lc : (Fin d →₀ ℕ) → ℝ) (P : Fin d →₀ ℕ)
(a : Fin d →₀ ℕ) : ℂ :=
w * ((lc a : ℝ) : ℂ) * (starRingEnd ℂ) (u a) * u (a - P)
/-! ## 3. The degree-preserving hops carry no flux -/
/-- The mode-exchange hop `α ↦ α − eⱼ + eᵢ`. -/
def shiftm (a : Fin d →₀ ℕ) (i j : Fin d) : Fin d →₀ ℕ :=
a - Finsupp.single j 1 + Finsupp.single i 1
/-- The amplitude of the mode-exchange hop: `√(αⱼ(αᵢ+1))` for `i ≠ j`, and the number
`αᵢ` for `i = j`. -/
def rcm (a : Fin d →₀ ℕ) (i j : Fin d) : ℝ :=
Real.sqrt ((a j : ℝ)) * Real.sqrt ((((a - Finsupp.single j 1 : Fin d →₀ ℕ) i : ℕ) : ℝ) + 1)
/-- The contribution of the mode-exchange hop `(i, j)`. -/
def mterm (u : (Fin d →₀ ℕ) → ℂ) (M : Fin d → Fin d → ℂ) (a : Fin d →₀ ℕ) (i j : Fin d) : ℂ :=
M i j * ((rcm a i j : ℝ) : ℂ) * (starRingEnd ℂ) (u a) * u (shiftm a i j)
/-! ## 4. The general quadratic recursion -/
/-- The shift of the pair hop `(i, j)`: `eᵢ + eⱼ`. -/
def pvec (i j : Fin d) : Fin d →₀ ℕ := Finsupp.single i 1 + Finsupp.single j 1
/-- The raising amplitude of the pair hop: `√((αᵢ+1)(αⱼ+1))` for `i ≠ j`, and
`√((αᵢ+1)(αᵢ+2))` for `i = j`. -/
def rcp (a : Fin d →₀ ℕ) (i j : Fin d) : ℝ :=
Real.sqrt ((a j : ℝ) + 1) * Real.sqrt ((((a + Finsupp.single j 1 : Fin d →₀ ℕ) i : ℕ) : ℝ) + 1)
/-- The lowering amplitude of the pair hop: `√(αᵢαⱼ)` for `i ≠ j`, and `√(αᵢ(αᵢ−1))` for
`i = j`. -/
def lcp (a : Fin d →₀ ℕ) (i j : Fin d) : ℝ :=
Real.sqrt ((a i : ℝ)) * Real.sqrt ((((a - Finsupp.single i 1 : Fin d →₀ ℕ) j : ℕ) : ℝ))
/-- **The general quadratic recursion**: a real diagonal `lam`, one-step amplitudes `w`,
pair creation/annihilation amplitudes `W`, and a mode-exchange matrix `M`. -/
def LadderRecQ (u : (Fin d →₀ ℕ) → ℂ) (lam : (Fin d →₀ ℕ) → ℝ) (w : Fin d → ℂ)
(W M : Fin d → Fin d → ℂ) (z : ℂ) : Prop :=
∀ a : Fin d →₀ ℕ,
((lam a : ℝ) : ℂ) * u a
+ ∑ i, ((starRingEnd ℂ) (w i) * ((rc1 a i : ℝ) : ℂ) * u (a + Finsupp.single i 1)
+ w i * ((lc1 a i : ℝ) : ℂ) * u (a - Finsupp.single i 1))
+ ∑ i, ∑ j, ((starRingEnd ℂ) (W i j) * ((rcp a i j : ℝ) : ℂ) * u (a + pvec i j)
+ W i j * ((lcp a i j : ℝ) : ℂ) * u (a - pvec i j))
+ ∑ i, ∑ j, (M i j * ((rcm a i j : ℝ) : ℂ) * u (shiftm a i j))
= z * u a
variable {lam : (Fin d →₀ ℕ) → ℝ} {w : Fin d → ℂ} {W M : Fin d → Fin d → ℂ} {z : ℂ}
/-! ## 5. The flux bound and the boundary mass -/
/-! ## 6. The criterion -/
end
end BookProof.CarlemanSimplex