AFTD Auto-Formalizing Theoretical Domains

thermally_reachable_subset_base_polytope

theoremverified

For any surjective block-assignment map proj : Fin n → Fin m and vectors γ, x : Fin n → ℝ, thermally_reachable_block_populations proj γ x is a subset of submodular_base_polytope (block_population_capacity proj γ x).

Statement

theorem thermally_reachable_subset_base_polytope {n m : ℕ} (proj : Fin n → Fin m) (hsurj : Function.Surjective proj) (γ x : Fin n → ℝ) : thermally_reachable_block_populations proj γ x ⊆ submodular_base_polytope (block_population_capacity proj γ x)
Source
arXiv:2610.35433:888
Verified
31 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 5
  • thermally_reachable_block_populationsGiven a block-assignment map proj from Fin n to Fin m and vectors γ, x : Fin n → ℝ, thermally_reachable_block_populations proj γ x is the…
  • block_population_capacityGiven a block-assignment map proj : Fin n → Fin m, vectors γ, x : Fin n → ℝ, and a subset S : Finset (Fin m), block_population_capacity…
  • submodular_base_polytopeThe base polytope B(f) of a set function f on Finset (Fin m) is the set of vectors z in ℝ^m such that the sum of z over every subset S is…
  • fiber_sum_vecGiven an assignment of n microstates to m energy levels, the fiber sum vector of a microstate distribution y has k-th component equal to…
  • thermally_reachable_microA microstate distribution x' on n energy levels is thermally reachable from an initial distribution x with respect to the Gibbs vector γ…
Used by 1
  • target_population_le_capacityFor any natural number n, target population vector y in stdSimplex ℝ (Fin n), Gibbs vector γ in stdSimplex ℝ (Fin n) with strictly…

Read back from the Lean

For any natural numbers $n$ and $m$, any surjective function $\text{proj} : \text{Fin } n \to \text{Fin } m$, and any vectors $\gamma, x : \text{Fin } n \to \mathbb{R}$, the set of thermally reachable block populations associated with $\text{proj}$, $\gamma$, and $x$ is a subset of the submodular base polytope associated with the block population capacity function for $\text{proj}$, $\gamma$, and $x$.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

The proof establishes that any thermally reachable block population p lies in the submodular base polytope of the block population capacity. First, for every subset of blocks S, the partial sum of p over S is bounded by the capacity of S because the capacity is defined as the supremum of such partial sums, which is non-empty and bounded above by 1. Second, the total block population sums to 1 because grouping the microscopic state x' by fibers and summing over all blocks reconstructs the full sum of x', which is 1 since x' lies in the standard simplex.

  1. To show that any thermally reachable block population belongs to the base polytope, it suffices to verify the capacity inequality for each block subset and check that the total block population sums to 1.

    Let p be a thermally reachable block population. By definition, p is obtained by summing the fiber coordinates of some microscopic state x' in the standard simplex that is thermally reachable from x. Membership in the submodular base polytope requires two conditions: the sum over each subset S is bounded by the capacity of S, and the sum over all coordinates equals 1.

    lines 3–6

      have _ := hsurj
      intro p hp
      rcases hp with ⟨x', hx'sim, hreach, rfl⟩
      constructor
  2. For each block subset S, the capacity constraint reduces to showing that the set of reachable partial sums over S is bounded above and contains the partial sum of p.

    The capacity of S is defined as the supremum of the partial sums over S across all thermally reachable block populations. By le_csSup, an element is bounded above by the supremum of a set provided that the set is bounded above and the element belongs to it.

    lines 7–9

    uses le_csSup

      · intro S
        dsimp [block_population_capacity]
        apply le_csSup
  3. The value 1 is an upper bound for the partial sum over S of any reachable block population.

    Any reachable block population q arises from a microscopic probability vector x'' in the standard simplex. Since each coordinate of x'' is non-negative, each block sum is non-negative by Finset.sum_nonneg. Hence, summing over S is at most the sum over all blocks by Finset.sum_le_univ_sum_of_nonneg. By Finset.sum_fiberwise, summing across all fibers reconstructs the total sum of x'', which is 1 because x'' lies in the standard simplex.

    lines 10–19

    uses Finset.sum_nonneg, Finset.sum_le_univ_sum_of_nonneg, Finset.sum_fiberwise

        · use 1
          rintro y ⟨q, hq, rfl⟩
          rcases hq with ⟨x'', hx''sim, _, rfl⟩
          dsimp [fiber_sum_vec]
          have hp_nonneg : ∀ k, 0 ≤ ∑ i with proj i = k, x'' i :=
            fun k => Finset.sum_nonneg (fun i _ => hx''sim.1 i)
          have h_le_univ : ∑ k ∈ S, (∑ i with proj i = k, x'' i) ≤ ∑ k, (∑ i with proj i = k, x'' i) :=
            Finset.sum_le_univ_sum_of_nonneg (fun k => hp_nonneg k)
          rw [Finset.sum_fiberwise Finset.univ proj x'', hx''sim.2] at h_le_univ
          exact h_le_univ
  4. The partial sum over S of p belongs to the set whose supremum defines the capacity.

    Because p is thermally reachable (as witnessed by x'), its partial sum over S is the image of p under the linear functional q ↦ ∑_{k ∈ S} q(k), which belongs to the image set by Set.mem_image_of_mem.

    line 20

    uses Set.mem_image_of_mem

        · exact Set.mem_image_of_mem _ ⟨x', hx'sim, hreach, rfl⟩
  5. The total block population sums to 1.

    Summing the block components of p = fiber_sum_vec proj x' over all blocks partitions Fin n by the fibers of proj. By Finset.sum_fiberwise, this equals the sum of the microscopic coordinates ∑ i, x' i, which equals 1 because x' lies in the standard simplex.

    lines 21–22

    uses Finset.sum_fiberwise

      · dsimp [fiber_sum_vec]
        rw [Finset.sum_fiberwise Finset.univ proj x', hx'sim.2]
le_csSup
In a conditionally complete lattice such as ℝ, every element of a set that is bounded above is at most the supremum of the set.
Finset.sum_nonneg
A finite sum of non-negative elements is non-negative.
Finset.sum_le_univ_sum_of_nonneg
For a function with non-negative values on a finite type, the sum over any subset is bounded by the sum over the entire type.
Finset.sum_fiberwise
Summing values fiberwise by grouping elements according to the value of a projection map and then summing over all fibers equals the original sum over the entire domain.
Set.mem_image_of_mem
If an element belongs to a set, its value under a function belongs to the image of that set under the function.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- Every thermally reachable block population vector satisfies the capacity constraints and belongs to the base polytope. -/theorem thermally_reachable_subset_base_polytope {n m : ℕ} (proj : Fin n → Fin m) (hsurj : Function.Surjective proj) (γ x : Fin n → ℝ) : thermally_reachable_block_populations proj γ x ⊆ submodular_base_polytope (block_population_capacity proj γ x) := by  have _ := hsurj  intro p hp  rcases hp with ⟨x', hx'sim, hreach, rfl⟩  constructor  · intro S    dsimp [block_population_capacity]    apply le_csSup    · use 1      rintro y ⟨q, hq, rfl⟩      rcases hq with ⟨x'', hx''sim, _, rfl⟩      dsimp [fiber_sum_vec]      have hp_nonneg : ∀ k, 0 ≤ ∑ i with proj i = k, x'' i :=        fun k => Finset.sum_nonneg (fun i _ => hx''sim.1 i)      have h_le_univ : ∑ k ∈ S, (∑ i with proj i = k, x'' i) ≤ ∑ k, (∑ i with proj i = k, x'' i) :=        Finset.sum_le_univ_sum_of_nonneg (fun k => hp_nonneg k)      rw [Finset.sum_fiberwise Finset.univ proj x'', hx''sim.2] at h_le_univ      exact h_le_univ    · exact Set.mem_image_of_mem _ ⟨x', hx'sim, hreach, rfl⟩  · dsimp [fiber_sum_vec]    rw [Finset.sum_fiberwise Finset.univ proj x', hx'sim.2]
See it in the graphDiscuss this resultChallenge it