Discounted occupation bounds before recurrence #
The reflected resolvent is an unnormalised time integral. Its uniform bound controls first moments of the cumulative occupations at an arbitrary positive rate.
References #
- J. Bertoin, Lévy Processes, Cambridge University Press, 1996, Chapter IV.
The ordinary time integral of the discounted reflected occupation kernel.
Equations
- BGW.StableTrees.supAuxOccupationTail q ε s f = ∫ (t : ℝ) in Set.Ioi ↑s, Real.exp (-q * t) * BGW.StableTrees.supOccKernel ε (BGW.StableTrees.supReflectedFrom 0 f t.toNNReal)
Instances For
The reflected occupation kernel is jointly measurable in real time and path.
The discounted occupation tail is measurable for every rate and cutoff.
A positive cutoff has an integrable discounted kernel on the time-path product.
The discounted occupation tail is integrable under the local stable law.
At zero the occupation tail has exactly the supOccupationResolvent mean,
without a rate factor.
Discounted occupation kernels are integrable on every individual path.
Every fixed-time occupation is integrable, using its deterministic cutoff bound.
Removing the discount on a finite interval costs at most exp (q T).
An arbitrary positive resolvent rate gives a cutoff-uniform first-moment growth bound.