Mean endpoint costs propagate through a finite kernel
ProvedDerivativeChainContinuation.kernel_mean_cost_boundcontinuation-20261026derivative-chainsgeometric-budgets
For nonnegative finite weights , a nonnegative row-stochastic kernel , and real endpoint costs , suppose for every pair. With and ,
Preamble
import Mathlib import Definitions.Def_DerivativeChainContinuation_FiniteNormMean import Definitions.Def_DerivativeChainPrivate_foundations open DerivativeChainPrivate set_option autoImplicit false set_option backward.isDefEq.respectTransparency false open Polynomial Matrix open scoped BigOperators set_option maxHeartbeats 1000000
Formal statement
theorem DerivativeChainContinuation.kernel_mean_cost_bound {ι κ : Type*} [Fintype ι] [Fintype κ]
(w : ι → ℝ) (K : ι → κ → ℝ) (x : ι → ℂ) (y : κ → ℂ)
(f : ι → ℝ) (g : κ → ℝ)
(hw : ∀ i, 0 ≤ w i) (hK : ∀ i a, 0 ≤ K i a) (hKs : ∀ i, ∑ a, K i a=1)
(hcost : ∀ i a, g a ≤ f i+‖y a-x i‖) :
(∑ a, (∑ i, w i*K i a)*g a) ≤ (∑ i, w i*f i)+(∑ i, w i*kernelCost K x y i) := by sorrySource