The convolution of two Poisson distributions is Poisson
ProvedFamousTheorems.poisson_convolution_poisson_7aThe convolution of two Poisson distributions is Poisson. For all rates ,
Equivalently, the sum of independent Poisson random variables with means and is Poisson with mean .
The proof is a computation with the binomial theorem, or with generating functions. This additivity makes the Poisson process well defined: counts in disjoint intervals are independent Poisson variables whose means add. It is used in queueing theory, in the modelling of rare events and in Poisson approximation.
Formalization note. Mathlib's ProbabilityTheory.poissonMeasure_conv_poissonMeasure. poissonMeasure r is the Poisson distribution with mean on , and Measure.conv is convolution of measures on the additive monoid .
Preamble
import Mathlib
Formal statement
namespace FamousTheorems
theorem poisson_convolution_poisson_7a (r₁ r₂ : NNReal) :
(ProbabilityTheory.poissonMeasure r₁).conv (ProbabilityTheory.poissonMeasure r₂) =
ProbabilityTheory.poissonMeasure (r₁ + r₂) := by sorry
end FamousTheoremsSource
Listed in Mathlib's curated theorem manifests (docs/1000.yaml); formalized in Mathlib as `ProbabilityTheory.poissonMeasure_conv_poissonMeasure`. Proof here reduces to that Mathlib result.