Provedadcgssupportingfeasibility
Both AdCGS outer averaging updates remain in the compact convex feasible set
For all , , , and , if , , , , and , then
What the notation means · or point at any symbol
- NNatural numbers
- The natural numbers 0, 1, 2, …. In Lean they start at 0. (
ℕ) - Set
- Set: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (
Set) - Convex
- Convex: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (
Convex) - x ∈ TVariables
- x, y, z range over EuclideanSpace ℝ (Fin n): the statement must hold for every choice. (
(x y z : EuclideanSpace ℝ (Fin n))) - ∈Element of
- x is a member of P. (
x ∈ P) - ≤At most
- 0 is less than or equal to b. (
0 ≤ b) - •Scalar multiplication
- y scaled by the number 1 − b. (
(1 - b) • y) - +Addition
- (1 − b) • y plus b • z. (
(1 - b) • y + b • z) - ∧And
- Both hold: (1 − b) • y + b • z ∈ P, and (tau/(1 + tau)) • x + (1/(1 + tau)) • z ∈ P. (
(1 - b) • y + b • z ∈ P ∧ (tau / (1 + t…) - -Subtraction
- 1 minus b. On ℕ, Lean's subtraction stops at 0: 2 − 5 = 0. (
1 - b) - EuclideanSpace
- EuclideanSpace: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (
EuclideanSpace) - RReal numbers
- The real numbers: every point on the number line. (
ℝ) - Fin
- Fin: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (
Fin) - f(x)Function application
- Fin applied to n. Lean writes f x; papers write f(x). (
Fin n) - a/bDivision
- tau divided by 1 + tau. In Lean, dividing by 0 gives 0. (
tau / (1 + tau))
theorem adcgs-outer-updates-preserve-feasibility (n : ℕ) (P : Set (EuclideanSpace ℝ (Fin n))) (hP : Convex ℝ P) (x y z : EuclideanSpace ℝ (Fin n)) (hx : x ∈ P) (hy : y ∈ P) (hz : z ∈ P) (b tau : ℝ) (hb0 : 0 ≤ b) (hb1 : b ≤ 1) (ht : 0 ≤ tau) :
(1 - b) • y + b • z ∈ P ∧ (tau / (1 + tau)) • x + (1 / (1 + tau)) • z ∈ PSection 2, immediately after equation (2), and Algorithm 1 lines 4–5. Convexity alone suffices; compactness is not needed for this local preservation lemma. This certifies the two affine updates given feasible inputs and admissible beta/tau, rather than assuming their output membership.
How it was proved
- Proved
- 30 September 2026
- By
- Ajax
- Checked by
- Lean's kernel (Lean 4.33.1, with Mathlib); the proof is signed and replays offline
- Axioms
- Classical.choice, Quot.sound, propext
- Proof
- 17 lines of Lean
Sign in to see the proofThe proof, its signed report and its history.