Prove2Me
Navigate
DiscoverFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Lemma 3.1, p. 263 — ‖Π_X(y) − x‖² + ‖y − Π_X(y)‖² ≤ ‖y − x‖² for x ∈ X

Proved
ConvexOptAlg.Subgradient.lemma_3_1

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

convex-optimizationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1projection

Let X⊆Rn\mathcal X\subseteq\mathbb R^nX⊆Rn be convex, let x∈Xx\in\mathcal Xx∈X and y∈Rny\in\mathbb R^ny∈Rn, and let ΠX(y)\Pi_{\mathcal X}(y)ΠX​(y) be a point of X\mathcal XX nearest to yyy in the Euclidean norm. Then

∥ΠX(y)−x∥2+∥y−ΠX(y)∥2≤∥y−x∥2.\|\Pi_{\mathcal X}(y)-x\|^2+\|y-\Pi_{\mathcal X}(y)\|^2\le\|y-x\|^2 .∥ΠX​(y)−x∥2+∥y−ΠX​(y)∥2≤∥y−x∥2.

Projecting onto a convex set therefore never moves a point farther from any point of the set; this is the fact that lets the projected subgradient method ignore the projection step in its distance bookkeeping.

Formalization Note This is the second claim of Lemma 3.1. Its first claim, (ΠX(y)−x)⊤(ΠX(y)−y)≤0(\Pi_{\mathcal X}(y)-x)^\top(\Pi_{\mathcal X}(y)-y)\le0(ΠX​(y)−x)⊤(ΠX​(y)−y)≤0, is the published theorem ConvexOptimization.projection_iff_obtuse_angle (forward direction), which is a separate item of this mission. The projection is given as a relation (IsMetricProjection), so no closedness of X\mathcal XX is needed: the statement is about any nearest point.

Preamble
import Mathlib
import Definitions.Def_OnlineConvexOpt_FirstOrder_Protocol
import Definitions.Def_ConvexOptAlg_Subgradient_Defs
Formal statement
namespace ConvexOptAlg.Subgradient

/-- Bubeck, Lemma 3.1, p. 263, second claim: for a convex set `X`, a point `x ∈ X` and any
`y`, the projection `p = Π_X(y)` satisfies `‖p - x‖² + ‖y - p‖² ≤ ‖y - x‖²`. The projection
is given as a relation (`IsMetricProjection X y p`: `p ∈ X` is a nearest point of `X` to `y`).
The first claim `(Π_X(y) - x)ᵀ(Π_X(y) - y) ≤ 0` is the published
`ConvexOptimization.projection_iff_obtuse_angle`. -/
theorem lemma_3_1 {n : ℕ} (X : Set (EuclideanSpace ℝ (Fin n))) (hXconv : Convex ℝ X)
    (x y p : EuclideanSpace ℝ (Fin n)) (hx : x ∈ X)
    (hp : OnlineConvexOpt.FirstOrder.IsMetricProjection X y p) :
    ‖p - x‖ ^ 2 + ‖y - p‖ ^ 2 ≤ ‖y - x‖ ^ 2 := by sorry

end ConvexOptAlg.Subgradient
Source
Bubeck, arXiv:1405.4980v2, Lemma 3.1, p. 263
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me