AFTD Auto-Formalizing Theoretical Domains

electrical_flow_cross_term

theoremverified

If F is a skew-symmetric flow on Fin k satisfying the voltage law d i j * F i j = h i - h j for all i ≠ j, and G is a skew-symmetric flow with divergence b (so ∑ j, G i j = b i for all i), then the cross term ∑ i, ∑ j with i < j, d i j * F i j * G i j equals ∑ i, b i * h i.

Statement

theorem electrical_flow_cross_term {k : ℕ} (d : Fin k → Fin k → ℝ) (F G : Fin k → Fin k → ℝ) (h : Fin k → ℝ) (b : Fin k → ℝ) (hF_skew : ∀ i j, F i j = -F j i) (hF_volt : ∀ i j, i ≠ j → d i j * F i j = h i - h j) (hG_skew : ∀ i j, G i j = -G j i) (hG_div : ∀ i, ∑ j, G i j = b i) : (∑ i : Fin k, ∑ j with i < j, d i j * F i j * G i j) = ∑ i, b i * h i
Source
arXiv:2611.05669, Appendix A
Verified
06 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 2
  • TotalPreorderA total preorder: a preorder where ≤ is total (complete). This is weaker than LinearOrder — it does NOT require antisymmetry or…
  • LinearOrder.toTotalPreorderEvery LinearOrder is a TotalPreorder.

Read back from the Lean

For any natural number $k$, given functions $d, F, G : \text{Fin } k \to \text{Fin } k \to \mathbb{R}$ and $h, b : \text{Fin } k \to \mathbb{R}$ satisfying: 1. $F$ is skew-symmetric: for all $i, j \in \text{Fin } k$, $F(i, j) = -F(j, i)$, 2. Ohm's law / potential difference condition: for all $i, j \in \text{Fin } k$ with $i \neq j$, $d(i, j) F(i, j) = h(i) - h(j)$, 3. $G$ is skew-symmetric: for all $i, j \in \text{Fin } k$, $G(i, j) = -G(j, i)$, 4. Divergence condition: for all $i \in \text{Fin } k$, $\sum_{j : \text{Fin } k} G(i, j) = b(i)$, then: $$\sum_{i : \text{Fin } k} \sum_{\substack{j : \text{Fin } k \\ i < j}} d(i, j) F(i, j) G(i, j) = \sum_{i : \text{Fin } k} b(i) h(i).$$

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 evaluates the cross term by discrete integration by parts. First, the voltage law replaces d(i,j) * F(i,j) with h(i) - h(j) on pairs i < j, splitting the sum into two parts. Swapping summation indices and using the skew-symmetry of G converts the subtracted term into a sum over j < i. Combining the two parts covers all j ≠ i (with the diagonal vanishing because G(i,i) = 0), so factoring out h(i) and using the divergence condition ∑_j G(i,j) = b(i) yields the inner product ∑_i b(i) * h(i).

  1. The restricted sum ∑_{i} ∑_{j: i < j} d(i,j) F(i,j) G(i,j) equals the sum over all pairs ∑_i ∑_j of the indicator term if i < j then (h(i) - h(j)) * G(i,j) else 0.

    The proof introduces hF_skew (line 3) and establishes lemma h1 (lines 4–13). By Finset.sum_filter, the sum over {j | i < j} is expressed as an indicator over all j. For pairs with i < j, we have i ≠ j by ne_of_lt, allowing the voltage condition hF_volt to replace d(i,j) * F(i,j) with h(i) - h(j). Line 14 rewrites the main goal with h1.

    lines 3–14

    uses Finset.sum_congr, Finset.sum_filter, ne_of_lt

      have _ := hF_skew
      have h1 : (∑ i : Fin k, ∑ j with i < j, d i j * F i j * G i j) =
          ∑ i, ∑ j, if i < j then (h i - h j) * G i j else 0 := by
        apply Finset.sum_congr rfl
        intro i _
        rw [Finset.sum_filter]
        apply Finset.sum_congr rfl
        intro j _
        split_ifs with hij
        · rw [hF_volt i j (ne_of_lt hij)]
        · rfl
      rw [h1]
  2. The double sum splits into the difference of two sums: (∑_i ∑_j if i < j then h(i) * G(i,j) else 0) - (∑_i ∑_j if i < j then h(j) * G(i,j) else 0).

    Lemma h2 (lines 15–24) distributes the subtraction (h(i) - h(j)) * G(i,j) = h(i) * G(i,j) - h(j) * G(i,j) across both the outer and inner sums using Finset.sum_sub_distrib and congruence Finset.sum_congr, verified pointwise by ring arithmetic. Line 25 rewrites the goal with h2.

    lines 15–25

    uses Finset.sum_sub_distrib, Finset.sum_congr

      have h2 : (∑ i : Fin k, ∑ j, if i < j then (h i - h j) * G i j else (0 : ℝ)) =
          (∑ i, ∑ j, if i < j then h i * G i j else 0) -
          (∑ i, ∑ j, if i < j then h j * G i j else 0) := by
        rw [← Finset.sum_sub_distrib]
        apply Finset.sum_congr rfl
        intro i _
        rw [← Finset.sum_sub_distrib]
        apply Finset.sum_congr rfl
        intro j _
        split_ifs <;> ring
      rw [h2]
  3. The second sum equals -∑_i ∑_j (if j < i then h(i) * G(i,j) else 0), so subtracting it turns the whole expression into ∑_i ((∑_j if i < j then h(i) * G(i,j) else 0) + (∑_j if j < i then h(i) * G(i,j) else 0)).

    Lemma h3 (lines 26–36) swaps the summation indices i and j in the second double sum using Finset.sum_comm, converting the condition i < j into j < i and the summand into h(i) * G(j,i). Applying the skew-symmetry of G (G(j,i) = -G(i,j)) pulls out a minus sign using Finset.sum_neg_distrib. On line 37, rewriting with h3 converts subtraction of a negative into addition via sub_neg_eq_add, and Finset.sum_add_distrib merges the outer sums over i.

    lines 26–37

    uses Finset.sum_comm, Finset.sum_neg_distrib, Finset.sum_congr, sub_neg_eq_add, Finset.sum_add_distrib

      have h3 : (∑ i : Fin k, ∑ j, if i < j then h j * G i j else (0 : ℝ)) =
          - ∑ i, ∑ j, if j < i then h i * G i j else 0 := by
        rw [Finset.sum_comm]
        rw [← Finset.sum_neg_distrib]
        apply Finset.sum_congr rfl
        intro i _
        rw [← Finset.sum_neg_distrib]
        apply Finset.sum_congr rfl
        intro j _
        rw [hG_skew j i]
        split_ifs <;> ring
      rw [h3, sub_neg_eq_add, ← Finset.sum_add_distrib]
  4. For each node i, the sum over j of the combined indicators equals b(i) * h(i).

    Lemma h4 (lines 38–51) fixes an index i and adds the two inner sums with Finset.sum_add_distrib. Sublemma hterm (lines 42–50) shows that for every j, the sum of the two indicators (if i < j then h(i)*G(i,j) else 0) + (if j < i then h(i)*G(i,j) else 0) equals h(i) * G(i,j): by lt_trichotomy on i and j, if i < j or j < i, one branch is active and the other is 0; if i = j, both branches are 0, which matches h(i) * G(i,i) because G(i,i) = 0 by skew-symmetry (hG_skew). Factoring out h(i) via Finset.mul_sum leaves h(i) * ∑_j G(i,j), which by the divergence condition hG_div evaluates to h(i) * b(i) = b(i) * h(i).

    lines 38–51

    uses Finset.sum_add_distrib, Finset.sum_congr, lt_trichotomy, if_pos, if_neg, asymm, add_zero, zero_add, Finset.mul_sum, mul_comm

      have h4 : ∀ i : Fin k, (∑ j, if i < j then h i * G i j else 0) +
          (∑ j, if j < i then h i * G i j else 0) = b i * h i := by
        intro i
        rw [← Finset.sum_add_distrib]
        have hterm : (∑ j, ((if i < j then h i * G i j else 0) + (if j < i then h i * G i j else 0))) =
            ∑ j, h i * G i j := by
          apply Finset.sum_congr rfl
          intro j _
          rcases lt_trichotomy i j with hij | rfl | hji
          · rw [if_pos hij, if_neg (asymm hij), add_zero]
          · have hG0 : G i i = 0 := by have := hG_skew i i; linarith
            simp [hG0]
          · rw [if_pos hji, if_neg (asymm hji), zero_add]
        rw [hterm, ← Finset.mul_sum, hG_div i, mul_comm]
  5. The total energy cross term equals ∑_i b(i) * h(i).

    Sum congruence Finset.sum_congr replaces each term in the outer sum over i with b(i) * h(i) using lemma h4, completing the proof.

    lines 52–54

    uses Finset.sum_congr

      apply Finset.sum_congr rfl
      intro i _
      exact h4 i
Finset.sum_congr
Two finite sums are equal if their index sets are equal and their summands agree on every element of the index set.
Finset.sum_filter
A sum over elements satisfying a predicate equals the sum over all elements with an if-then-else indicator.
ne_of_lt
In a preorder, if a < b then a ≠ b.
Finset.sum_sub_distrib
The sum of a difference equals the difference of the sums: ∑ (f - g) = ∑ f - ∑ g.
Finset.sum_comm
The order of summation in a double finite sum can be interchanged: ∑_i ∑_j f(i,j) = ∑_j ∑_i f(i,j).
Finset.sum_neg_distrib
Negation distributes over finite summation: ∑ (-f) = -∑ f.
sub_neg_eq_add
Subtracting a negative quantity is equivalent to addition: a - (-b) = a + b.
Finset.sum_add_distrib
The sum of a sum equals the sum of the individual sums: ∑ (f + g) = ∑ f + ∑ g.
lt_trichotomy
In any linear order, exactly one of a < b, a = b, or b < a holds.
if_pos
If a proposition c is true, the conditional expression (if c then t else e) evaluates to t.
if_neg
If a proposition c is false, the conditional expression (if c then t else e) evaluates to e.
asymm
An asymmetric relation satisfies r(a, b) → ¬r(b, a); in particular, a < b implies ¬(b < a).
add_zero
Adding zero on the right leaves an element unchanged: a + 0 = a.
zero_add
Adding zero on the left leaves an element unchanged: 0 + a = a.
Finset.mul_sum
A factor distributes from the left over a finite sum: a * ∑ f_i = ∑ (a * f_i).
mul_comm
Multiplication is commutative: a * b = b * a.

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

/-- The cross term in the energy of a potential-induced flow and a feasible flow equals the inner product of potentials and demands. -/theorem electrical_flow_cross_term {k : ℕ} (d : Fin k → Fin k → ℝ) (F G : Fin k → Fin k → ℝ) (h : Fin k → ℝ) (b : Fin k → ℝ) (hF_skew : ∀ i j, F i j = -F j i) (hF_volt : ∀ i j, i ≠ j → d i j * F i j = h i - h j) (hG_skew : ∀ i j, G i j = -G j i) (hG_div : ∀ i, ∑ j, G i j = b i) : (∑ i : Fin k, ∑ j with i < j, d i j * F i j * G i j) = ∑ i, b i * h i := by  have _ := hF_skew  have h1 : (∑ i : Fin k, ∑ j with i < j, d i j * F i j * G i j) =      ∑ i, ∑ j, if i < j then (h i - h j) * G i j else 0 := by    apply Finset.sum_congr rfl    intro i _    rw [Finset.sum_filter]    apply Finset.sum_congr rfl    intro j _    split_ifs with hij    · rw [hF_volt i j (ne_of_lt hij)]    · rfl  rw [h1]  have h2 : (∑ i : Fin k, ∑ j, if i < j then (h i - h j) * G i j else (0 : ℝ)) =      (∑ i, ∑ j, if i < j then h i * G i j else 0) -      (∑ i, ∑ j, if i < j then h j * G i j else 0) := by    rw [← Finset.sum_sub_distrib]    apply Finset.sum_congr rfl    intro i _    rw [← Finset.sum_sub_distrib]    apply Finset.sum_congr rfl    intro j _    split_ifs <;> ring  rw [h2]  have h3 : (∑ i : Fin k, ∑ j, if i < j then h j * G i j else (0 : ℝ)) =      - ∑ i, ∑ j, if j < i then h i * G i j else 0 := by    rw [Finset.sum_comm]    rw [← Finset.sum_neg_distrib]    apply Finset.sum_congr rfl    intro i _    rw [← Finset.sum_neg_distrib]    apply Finset.sum_congr rfl    intro j _    rw [hG_skew j i]    split_ifs <;> ring  rw [h3, sub_neg_eq_add, ← Finset.sum_add_distrib]  have h4 : ∀ i : Fin k, (∑ j, if i < j then h i * G i j else 0) +      (∑ j, if j < i then h i * G i j else 0) = b i * h i := by    intro i    rw [← Finset.sum_add_distrib]    have hterm : (∑ j, ((if i < j then h i * G i j else 0) + (if j < i then h i * G i j else 0))) =        ∑ j, h i * G i j := by      apply Finset.sum_congr rfl      intro j _      rcases lt_trichotomy i j with hij | rfl | hji      · rw [if_pos hij, if_neg (asymm hij), add_zero]      · have hG0 : G i i = 0 := by have := hG_skew i i; linarith        simp [hG0]      · rw [if_pos hji, if_neg (asymm hji), zero_add]    rw [hterm, ← Finset.mul_sum, hG_div i, mul_comm]  apply Finset.sum_congr rfl  intro i _  exact h4 i
See it in the graphDiscuss this resultChallenge it