Modern Stochastics: Theory and Applications

This paper introduces a novel stochastic partial differential equation (SPDE) driven by space-only fractional Lévy noise ${\dot{L}_{x}^{d}}$ with $d\in (0,1/2)$, governed by a second-order elliptic operator with piecewise constant coefficients. We establish a new framework for analyzing diffusion in heterogeneous media, proving the existence, uniqueness, and equivalence of mild and weak solutions for the resulting non-Gaussian random field. A key original contribution is the explicit characterization of the covariance kernel in terms of transmission-reflection paths induced by the interface. We provide sharp estimates for the sample path regularity, proving that the solution is almost surely Hölder continuous of order $\gamma \lt d$ in space and $\beta \lt \frac{2d+1}{4}$ in time. Our results demonstrate that the fractional order of the noise robustly dictates path regularity, which remains preserved despite the singular transmission conditions and the jump structure of the driving noise.

We study dynamical Borel–Cantelli properties for the left shift on the binary symbolic space equipped with a stationary determinantal measure generated by a Toeplitz convolution kernel. For sequences of cylinder sets whose measures have a divergent sum, we give sufficient conditions ensuring infinitely many visits almost surely. Under exponential decay of the Fourier coefficients and a uniform nondegeneracy assumption, we also establish the strong Borel–Cantelli property.

We introduce a dynamic cutting approach for the numerical approximation of Lévy-driven stochastic differential equations. The key idea is to remove small jumps according to a time-dependent threshold, so that the retained jumps form a time-inhomogeneous compound Poisson process. We derive ${L^{p}}$-strong convergence rates for the adjusted Euler scheme. Numerical experiments at matched computational cost compare the dynamic cutting scheme with the classical Asmussen–Rosiński truncation and demonstrate consistently smaller strong errors when the time-dependent jump coefficient has a singularity at $t=0$.

This paper investigates fractional Riesz–Bessel equations with random initial conditions that exhibit either classical or cyclic long-range dependence. It studies zoom-in asymptotics for the corresponding solutions and establishes multiscaling limit theorems. It is known that for similar problems, non-degenerate multiscaling limits may not exist in general. The paper develops a kernel-smoothing approach for these equations and obtains non-degenerate limit fields under suitable normalisation and rescaling. It proves that the kernel-smoothed solutions converge weakly to Gaussian random fields, which are non-stationary in both time and space. Their stochastic integral representations and covariance functions are derived. The paper also analyses the regularity and dependence structure of the limit fields. In particular, under appropriate general assumptions on the smoothing kernel, the limits exhibit long-range dependence in time and short-range dependence in space. Numerical examples for the case of Matérn-type kernels are provided to illustrate the theoretical results.

We consider an extended variant of the classical coupon collector’s problem with an infinite number of collections. An arriving coupon is placed in the rth collection, $r\ge 0$, if r is the smallest index such that the corresponding collection still does not have a coupon of this type. We derive distributional limit theorems for the number of empty spots in different collections at the time when the 0th collection was completed, as well as after some delay. We also obtain the joint limiting distribution for completion times of different collections. All main results are given in an ultimate infinite-dimensional form in the sense of distributional convergence in ${\mathbb{R}^{\infty }}$. The main tool in the proofs is convergence of specially constructed point processes.

We investigate the fractional Riemann-Liouville integral of a general continuous Gaussian process, focusing first on the functional weak convergence of suitably rescaled processes. Building on these results and recent theoretical advances, we deduce functional large deviation principles. For small times, the asymptotic behavior depends solely on the covariance of the underlying process at zero, while for large times, appropriate rescaling of the covariance is required and additional regularity assumptions must be imposed. A central aspect of our work is the explicit characterization of the reproducing kernel Hilbert spaces of the fractionally integrated processes, including concrete formulas for the related norms. As a notable special case, when the underlying Gaussian process is Gauss-Markov, the reproducing kernel Hilbert spaces and the covariance structure can be described explicitly, providing precise insights into the corresponding rate functions.

We study a level-dependent PH/PH/1 queue with an infinite waiting room and derive explicit density formulas for the associated Matrix Mittag-Leffler (MML) random variables. The motivation arises from the use of Phase Type (PH) and Mittag-Leffler type distributions in modeling heavy-tailed and non-Markovian arrival and service processes in queueing systems, where classical exponential assumptions may fail to capture realistic behavior. Our goal is to obtain explicit density expressions for the MML random variables associated with the truncated queueing systems PH/M/$1\sim n$ and M/PH/$1\sim n$, where n denotes the truncation level at which the queueing process is absorbed. We first analyze PH/M/1 queues in which the inter-arrival times are PH-distributed, focusing specifically on exponential and Erlang cases, and derive closed-form MML densities. We then consider M/PH/1 queues with PH-distributed service times and obtain analogous explicit density formulas. Several examples with PH-generators of different orders are presented to illustrate the results.