Finite interval-sum bound from adjacent unit comparisons
ProvedAvramDividend.Classical.chat07_adjacent_interval_integral_bound_finavram-dividend-classicalfinite-summationintegrabilityinterval-induction
Let f and g be real-valued functions interval integrable over every finite real interval and let c be real. Suppose for every natural n, the interval integral of f from n to n+1 is at most c times the interval integral of g from n+1 to n+2. Then for every natural N, the integral of f from 0 to N is at most c times the integral of g from 1 to N+1. The proof is a finite induction using additivity of interval integrals, with no positivity hypothesis or infinite-series machinery. This supports the Avram BV scale-function weighted-derivative integrability programme.
Preamble
import Mathlib.MeasureTheory.Integral.IntervalIntegral.Basic open MeasureTheory intervalIntegral open scoped NNReal ENNReal
Formal statement
theorem AvramDividend.Classical.chat07_adjacent_interval_integral_bound_fin
(f g : ℝ → ℝ) (c : ℝ)
(hf : ∀ a b : ℝ, IntervalIntegrable f volume a b)
(hg : ∀ a b : ℝ, IntervalIntegrable g volume a b)
(hstep : ∀ n : ℕ,
(∫ x in (n : ℝ)..((n : ℝ) + 1), f x) ≤
c * (∫ x in ((n : ℝ) + 1)..((n : ℝ) + 2), g x))
(n : ℕ) :
(∫ x in (0 : ℝ)..(n : ℝ), f x) ≤
c * (∫ x in (1 : ℝ)..((n : ℝ) + 1), g x) := by sorry