theorem
Determinize.Proof.RewardModel.shift_iSup
(r : ℚ)
(μ : ℕ → MeasureTheory.Measure ℝ)
(mono : Monotone μ)
:
theorem
Determinize.Proof.RewardModel.outputAt_equation
(model : Spec.RewardModel.Model)
(i : Fin model.size)
:
model.outputAt i = match model.kind i with
| Spec.FiniteModel.StateKind.returned b => MeasureTheory.Measure.dirac ↑b
| Spec.FiniteModel.StateKind.rejected => 0
| Spec.FiniteModel.StateKind.transient =>
(List.map
(fun (e : Spec.RewardModel.Edge model.size) =>
ENNReal.ofReal ↑e.probability • Spec.RewardModel.shift e.reward (model.outputAt e.target))
(model.edges i)).sum
theorem
Determinize.Proof.RewardModel.outputAt_mass_le_one
(model : Spec.RewardModel.Model)
(i : Fin model.size)
:
instance
Determinize.Proof.RewardModel.instIsFiniteMeasureRealOutputAt
(model : Spec.RewardModel.Model)
(i : Fin model.size)
:
MeasureTheory.IsFiniteMeasure (model.outputAt i)