On an asymptotic distribution connected with a loss queueing system
In this paper we consider the problem of estimating the intensity of a spatial homogeneous Poisson process if a part of the observations (quadrat counts) is censored. The actual problem has occurred during a court case when one of the authors was a referee for the defense.
An exponential inequality for Choquet expectation is discussed. We also obtain a strong law of large numbers based on Choquet expectation. The main results of this paper improve some previous results obtained by many researchers.
In this paper we derive conditions upon the nonnegative random variable under which the inequality holds for a fixed nonnegative constant and for any absolutely continuous function . Taking into account the characterization of a Gamma distribution we consider the functional and establishing some of its properties we show that and that iff the random variable has a Gamma distribution.
In a recent paper, Neuts, Rauschenberg and Li [10] examined, by computer experimentation, four different procedures to randomly partition the interval [0,1] into m intervals. The present paper presents some new theoretical results on one of the partitioning schemes. That scheme is called Random Interval (RI); it starts with a first random point in [0,1] and places the kth point at random in a subinterval randomly picked from the current k subintervals (1
An example of a nonzero σ-finite Borel measure μ with everywhere dense linear manifold of admissible (in the sense of invariance) translation vectors is constructed in the Hilbert space ℓ₂ such that μ and any shift of μ by a vector are neither equivalent nor orthogonal. This extends a result established in [7].
We show that in the space C[-1,1] there exists an orthogonal algebraic polynomial basis with optimal growth of degrees of the polynomials.
The paper is, for the most part, devoted to a survey of the analytical properties of generalized convolution algebras and their realizations. This issue appears to be the state of the art until now because intensive research on the generalized convolution and the related models still persists.