Abstract
The Bernoulli sieve is the infinite ‘balls-in-boxes’ occupancy scheme with random frequencies , where
are independent copies of a random variable W taking values in (0, 1). Assuming that the number of balls equals n, let L
n
denote the number of empty boxes within the occupancy range. The paper proves that, under a regular variation assumption, L
n
, properly normalized without centring, weakly converges to a functional of an inverse stable subordinator. Proofs rely upon the observation that (log P
k
) is a perturbed random walk. In particular, some results for general perturbed random walks are derived. The other result of the paper states that whenever L
n
weakly converges (without normalization) the limiting law is mixed Poisson.
Notes
1. When the law of ξ is d-lattice, the standard form of the key renewal theorem proves (Equation13) with the limit taken over t ∈ dN. Noting that in the case Eξ = ∞
, for any y ∈ R, and using Feller's classical approximation argument (see p. 361–362 in [Citation7]) lead to (Equation13
).