Documentation

FirstOrderMethodsOptimization_Beck_2017.Chap06.Example_6_61

Statements

Source excerpts; proofs omitted. recall restates an existing declaration; #check records an expression, not its inferred type.

Example 6.61 (1): for a convex set C, the half squared distance function x ↦ (1 / 2) d_C(x)^2 = (infDist x C)^2 / 2 is globally 1-smooth.

recall half_sq_infDist_is_l_smooth

-- Proof sketch: combine the indicator/projection identification
-- `prox_extendedIndicator_eq_projection_mapping` with the singleton projection theorem
-- `projection_mapping_eq_singleton_of_nonempty_closed_convex`.

Lean source

Source context
universe u
variable {E : Type u} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [CompleteSpace E]

Example 6.61 (2): for a nonempty closed convex set C, the gradient of x ↦ (1 / 2) d_C(x)^2 is x - P_C(x), where the chapter's source-facing proximal notation identifies prox_{δ_C}(x) with the same singleton by prox_extendedIndicator_eq_singleton_metricProjection.

recall gradient_half_sq_infDist_eq_sub_metricProjection

Lean source

Source context
universe u
variable {E : Type u} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [CompleteSpace E]

Module source

theorem prox_extendedIndicator_eq_singleton_metricProjection {E : Type u} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [CompleteSpace E] (C : Set E) (hC_nonempty : C.Nonempty) (hC_closed : IsClosed C) (hC_convex : Convex ℝ C) (x : E) :
prox[δ_ C] x = {projectionPoint C hC_nonempty hC_closed hC_convex x}

For a nonempty closed convex set, the proximal mapping of the indicator δ_C is the singleton containing the metric projection P_C(x).