Variable-coefficient Grönwall estimate #
The scalar comparison argument below allows an integrable, time-dependent coefficient. This is the form needed when the linear operator is merely measurable: replacing its integral by a uniform bound would lose information.
theorem
AVenhance.Infra.ODE.Gronwall.integral_gronwall_endpoint
{a b C : ℝ}
(hab : a ≤ b)
{c z : ℝ → ℝ}
(hc : IntervalIntegrable c MeasureTheory.volume a b)
(hcz : IntervalIntegrable (fun (t : ℝ) => c t * z t) MeasureTheory.volume a b)
(hc_nonneg : ∀ t ∈ Set.Icc a b, 0 ≤ c t)
(hineq : ∀ t ∈ Set.Icc a b, z t ≤ C + ∫ (s : ℝ) in a..t, c s * z s)
:
theorem
AVenhance.Infra.ODE.integral_gronwall_bound
{a b C : ℝ}
(hab : a ≤ b)
{c z : ℝ → ℝ}
(hc : IntervalIntegrable c MeasureTheory.volume a b)
(hcz : IntervalIntegrable (fun (t : ℝ) => c t * z t) MeasureTheory.volume a b)
(hc_nonneg : ∀ t ∈ Set.Icc a b, 0 ≤ c t)
(hineq : ∀ t ∈ Set.Icc a b, z t ≤ C + ∫ (s : ℝ) in a..t, c s * z s)
(t : ℝ)
:
Integral Grönwall inequality with a nonnegative integrable time-dependent coefficient. The additive constant may already include an integral of an inhomogeneous forcing term.
theorem
AVenhance.Infra.ODE.LinearODEData.IsSolution.norm_le_gronwall
{E : Type u_1}
[NormedAddCommGroup E]
[NormedSpace ℝ E]
{a b : ℝ}
{hab : a ≤ b}
(D : LinearODEData a b hab)
(u : C(↑(Set.Icc a b), E))
(hu : D.IsSolution u)
(t : ℝ)
:
The norm of the unique integral solution is controlled by the L¹ norms of the forcing and operator coefficient.