Documentation

FirstOrderMethodsOptimization_Beck_2017.Chap15.Definition_15_6

Statements

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

Definition 15.6: the linear-composite objective x ↦ f₁(x) + f₂(Ax) is the Chapter 10 composite objective specialized to g := f₂ ∘ A.

recall composite_model_objective

Lean source

Source context
universe u v
variable {X : Type u} {Y : Type v}
variable [AddCommMonoid X] [Module ℝ X]
variable [AddCommMonoid Y] [Module ℝ Y]
recall composite_model_objective_apply

Lean source

Source context
universe u v
variable {X : Type u} {Y : Type v}
variable [AddCommMonoid X] [Module ℝ X]
variable [AddCommMonoid Y] [Module ℝ Y]

The proper/closed/convex standing assumptions in Definition 15.6 are already the Chapter 15 owner IsADMMConvexObjectivePair; the linear-composite file adds no second regularity owner.

recall IsADMMConvexObjectivePair

Lean source

Source context
universe u v
variable {X : Type u} {Y : Type v}
variable [TopologicalSpace X] [AddCommMonoid X] [Module ℝ X]
variable [TopologicalSpace Y] [AddCommMonoid Y] [Module ℝ Y]

Module source

@[simp]
theorem composite_model_objective_comp_apply {X : Type u} {Y : Type v} [AddCommMonoid X] [Module ℝ X] [AddCommMonoid Y] [Module ℝ Y] (f₁ : X → EReal) (f₂ : Y → EReal) (A : X →ₗ[ℝ] Y) (x : X) :
composite_model_objective f₁ (f₂ ∘ ⇑A) x = f₁ x + f₂ (A x)

Evaluating the Chapter 10 composite owner at the linear-composite specialization recovers the textbook formula f₁(x) + f₂(Ax).