Weighted connectivity identities
OpenMCA.connectivity_weightedThis theorem is the diagonal-elasticity specialization of the connectivity theorem, corresponding to equations (18) and (19) of the source. Let e,v be parameter and rate coordinates. Assume ve is diagonal, e_i ve(i,i)=v_i, and veInv(i,i)=e_i/v_i, together with the same chain-rule and inverse hypotheses as the general theorem. Then the component sums are -1 on the diagonal and 0 off the diagonal, and the total-rate weighted connectivity sum is zero. Real division is total.
import Mathlib
open scoped BigOperators
open Matrix
namespace MCA
theorem connectivity_weighted
{n m : ℕ}
(e v : Fin n → ℝ)
(Fv : Matrix (Fin m) (Fin n) ℝ)
(vS : Matrix (Fin n) (Fin m) ℝ)
(ve : Matrix (Fin n) (Fin n) ℝ)
(Se : Matrix (Fin m) (Fin n) ℝ)
(W : Matrix (Fin n) (Fin n) ℝ)
(veInv : Matrix (Fin n) (Fin n) ℝ)
(fsInv : Matrix (Fin m) (Fin m) ℝ)
(hfs_left : fsInv * (Fv * vS) = 1)
(hfs_right : (Fv * vS) * fsInv = 1)
(hve_left : veInv * ve = 1)
(hve_right : ve * veInv = 1)
(h_chain : (Fv * vS) * Se + Fv * ve = 0)
(h_total : W = vS * Se + ve)
(h_offdiag : ∀ i j, i ≠ j → ve i j = 0)
(h_diag_scale : ∀ i, e i * ve i i = v i)
(h_inv_entries : ∀ i, veInv i i = e i / v i) :
(∀ i j, ∑ k, (e k / v k) * Se i k * vS k j = if i = j then -1 else 0) ∧
(∀ i j, ∑ k, (e k / v k) * W i k * vS k j = 0) := by sorry
end MCARead-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For all n,m including zero, functions e,v:Fin n -> R and matrices Fv:m x n, vS:n x m, ve:n x n, Se:m x n, W:n x n, veInv:n x n, fsInv:m x m, assume fsInv(Fv vS)=I_m, (Fv vS)fsInv=I_m, veInv ve=I_n, ve veInv=I_n, ve(i,j)=0 for i≠j, e(i)ve(i,i)=v(i), veInv(i,i)=e(i)/v(i), (Fv vS)Se+Fv ve=0, and W=vS Se+ve. Then sum_k (e(k)/v(k))Se(i,k)vS(k,j) is -1 if i=j and 0 otherwise, while sum_k (e(k)/v(k))W(i,k)vS(k,j)=0. Division is total and empty finite indices are allowed.