The exponential is the only continuous memoryless distribution (p.21)
ProvedQueueingFundamentals.Foundations.memoryless_only_exponentialcharacterizationexponential-distributionmemorylessp2o-batch-b23bp2o-gran-per-chapterp2o-plan-bookp2o-v1
Let be a random variable with values in almost surely, whose distribution is continuous ( for every real ). Suppose has the memoryless property (1.17): for all ,
Then there is such that is exponentially distributed with rate , i.e. for .
Together with (1.17) this characterizes the exponential distribution among continuous distributions on .
Formalization Note "Continuous distribution" is read as a continuous distribution function (no atoms). The conditional probability is Mathlib's cond, which is the zero measure when the conditioning event has probability zero; the hypothesis then forces that case not to occur.
Preamble
import Mathlib open MeasureTheory ProbabilityTheory
Formal statement
namespace QueueingFundamentals.Foundations
/-- p.21: the exponential distribution is the only continuous distribution on `[0, ∞)` with the
memoryless property (1.17). -/
theorem memoryless_only_exponential {Ω : Type*} [MeasurableSpace Ω] (μ : Measure Ω)
[IsProbabilityMeasure μ] (T : Ω → ℝ) (hTm : Measurable T)
(hcont : ∀ a : ℝ, μ {ω | T ω = a} = 0) (hnonneg : μ {ω | T ω < 0} = 0)
(hmem : ∀ t₀ t₁ : ℝ, 0 ≤ t₀ → t₀ ≤ t₁ →
cond μ {ω | t₀ ≤ T ω} {ω | T ω ≤ t₁} = μ {ω | 0 ≤ T ω ∧ T ω ≤ t₁ - t₀}) :
∃ lam : ℝ, 0 < lam ∧ μ.map T = expMeasure lam := by sorry
end QueueingFundamentals.Foundations
Source
Gross, Shortle, Thompson & Harris, Fundamentals of Queueing Theory, 4th ed., Wiley 2008, DOI 10.1002/9781118625651, p.21, characterization of the exponential distribution following Eq. (1.17) (unnumbered)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.