Theorem 11 — limits of γ–tilted convex functions
OpenTiltedCvx.Supplement.theorem_11convex-analysisgeneralized-convexitynonsmooth-analysisoptimizationp2o-gran-per-chapterp2o-plan-bookp2o-strengthenedp2o-v1
Let be a non-empty convex set and . Let be proper, lower semicontinuous, –tilted convex functions that converge pointwise to finite values on a dense subset . Then there is a continuous function such that
- uniformly on compact subsets of ;
- the function is –tilted convex.
Moreover, if there is a lower semicontinuous function with pointwise, then
and so it is –tilted convex.
This extends the classical stability of convex functions under pointwise limits to the class .
Formalization Note The members of are elements of the published classF γ. The envelopes are lscEnv (ri K) f and lscEnv K g; lower semicontinuity of on is LowerSemicontinuousOn g K, and the convergence is pointwise on in .
Differences from the source:
- generalization: functions on ℝ^d → functions on a finite-dimensional real inner product space E; reason: every such E is isometric to ℝ^d and ℝ^d = EuclideanSpace ℝ (Fin d) is an instance, so the statement specializes to the paper's
- encoding: ℝ̄ = ℝ ∪ {±∞}, dom(f) = {f finite} → EReal and the published TiltedCvx.effDom; reason: Lean convention; same instances
- encoding: D is a dense subset of K → D ⊆ K ⊆ closure D; reason: the meaning of density in K
- encoding: a continuous f : ri(K) → ℝ → f : E → ℝ with ContinuousOn f (ri K), read only on ri K; reason: values off ri(K) never enter (lscEnv (ri K) and the uniform bounds read only ri K)
- encoding: fₙ → f uniformly on compact subsets of ri(K) → for every compact C ⊆ ri K and ε > 0, eventually f x − ε ≤ fₙ x ≤ f x + ε on C (in EReal); reason: uniform convergence of ℝ̄-valued fₙ to a real f; it forces fₙ to be finite on C eventually, as |fₙ − f| ≤ ε requires
- reading: lsc(f) is γ–tilted convex (and so is lsc(g)) → lsc(f) ∈ classF γ and lsc(g) ∈ classF γ: proper, lsc and γ–tilted convex; reason: Definition 3 defines γ–tilted convexity for proper lsc functions; the envelope is lsc and is proper here, so the statement inside the definition's scope is the non-weaker reading
Preamble
import Mathlib import Definitions.Def_TiltedCvx_classF import Definitions.Def_TiltedCvx_Supplement_lscEnv open Filter Topology
Formal statement
namespace TiltedCvx.Supplement
/-- **Theorem 11 (limits of γ–tilted convex functions).** Pedramfar–Khanna–Quinn–Aggarwal,
arXiv:2611.03472, Theorem 11, p.21. Let `K ⊆ ℝ^d` be a non-empty convex set and `γ ∈ (0, 1]`.
Let `fₙ : ℝ^d → ℝ̄` be proper, lsc, and γ–tilted convex functions that converge pointwise to
finite values on a dense subset `D ⊆ K`. Then there is a continuous function `f : ri(K) → ℝ` such
that (i) `fₙ → f` uniformly on compact subsets of `ri(K)`; (ii) the function `lsc(f) : ℝ^d → ℝ̄` is
γ–tilted convex. Moreover, if there is an lsc function `g : K → ℝ̄` such that `fₙ → g` pointwise,
then `lsc(g) = lsc(f)` and so it is γ–tilted convex.
**Formalization Note** "Proper, lsc, and γ–tilted convex" is membership in the published
`classF γ`; the conclusions "is γ–tilted convex" are stated as membership in `classF γ`, since
Definition 3 applies to proper lsc functions. "`D` dense in `K`" is `D ⊆ K ⊆ closure D`.
`f : ri(K) → ℝ` is a real function on `E` read only on `ri(K) = intrinsicInterior ℝ K`.
Uniform convergence on a compact `C ⊆ ri(K)`: for every `ε > 0`, eventually
`f x − ε ≤ fₙ x ≤ f x + ε` for all `x ∈ C` (so `fₙ` is eventually finite on `C`).
`lsc(f)` is `lscEnv (ri K) f` (the envelope of `f` extended by `+∞` off `ri(K)`), and `lsc(g)` for
`g : K → ℝ̄` is `lscEnv K g`; lower semicontinuity of `g` as a function on `K` is
`LowerSemicontinuousOn g K`, and the pointwise convergence `fₙ → g` is on `K`, in `EReal`. -/
theorem theorem_11 {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E]
[FiniteDimensional ℝ E] (K : Set E) (hKne : K.Nonempty) (hKc : Convex ℝ K)
(γ : ℝ) (hγ0 : 0 < γ) (hγ1 : γ ≤ 1) (fn : ℕ → E → EReal)
(hfn : ∀ n, fn n ∈ (TiltedCvx.classF γ : Set (E → EReal)))
(D : Set E) (hDK : D ⊆ K) (hKD : K ⊆ closure D)
(hconv : ∀ x ∈ D, ∃ L : ℝ, Tendsto (fun n => fn n x) atTop (𝓝 (L : EReal))) :
∃ f : E → ℝ, ContinuousOn f (intrinsicInterior ℝ K) ∧
(∀ C ⊆ intrinsicInterior ℝ K, IsCompact C → ∀ ε : ℝ, 0 < ε →
∀ᶠ n in atTop, ∀ x ∈ C,
(((f x - ε : ℝ) : EReal) ≤ fn n x ∧ fn n x ≤ ((f x + ε : ℝ) : EReal))) ∧
lscEnv (intrinsicInterior ℝ K) (fun x => ((f x : ℝ) : EReal)) ∈
(TiltedCvx.classF γ : Set (E → EReal)) ∧
(∀ g : E → EReal, LowerSemicontinuousOn g K →
(∀ x ∈ K, Tendsto (fun n => fn n x) atTop (𝓝 (g x))) →
lscEnv K g = lscEnv (intrinsicInterior ℝ K) (fun x => ((f x : ℝ) : EReal)) ∧
lscEnv K g ∈ (TiltedCvx.classF γ : Set (E → EReal))) := by sorry
end TiltedCvx.Supplement
Source
Pedramfar–Khanna–Quinn–Aggarwal, Tilted Convexity and an Exact Offline–Online Separation, arXiv:2611.03472, Theorem 11, p.21