Lévy, Ornstein-Uhlenbeck, and subordination: spectral vs. jump description (Q2487832)

Brownian motions are a single example of the much more general familiy of Lévy motions -- i.e. motions performed by stochastic processes with stationary and independent increments. The main features distinguishing Brownian motion from all other Lévy motions are the continuity of sample-path trajectories, and the fact of possessing finite variance. All other Lévy processes are pure-jump processes -- their sample-path trajectories are purely discontinuous-- and have infinite variance [\textit{O. E. Barndorff-Nielsen, T. Mikosch} and \textit{S. Resnic} (eds), Lévy processes. Theory and applications. Boston: Birkhäuser) (2001; Zbl 0961.00012); \textit{M. F. Schlesinger, G. M. Zaslavsky} and \textit{U. Frisch} (eds), Lévy flights and related topics. Lecture Notes in Physics. 450. Berlin: Springer-Verlag (1995; Zbl 0823.00016); \textit{K. Sato}, Lévy processes and infinitely divisible distributions. Cambridge Studies in Advanced Mathematics. 68. Cambridge: Cambridge University Press (1999; Zbl 0973.60001)]. When analyzing non-Brownian Lévy motions, the common practice is to use either spectral methods, transforming to Fourier or Laplace space and translating probabilistic problems into analytic ones, or fractional-calculus methods, generalizing the diffusion and Fokker-Planck equations to superdiffusive behaviour [\textit{R. Metzler} and \textit{J. Klafter}, Phys. Rep. 339, No. 1, 1--77 (2000; Zbl 0984.82032)]. However, the spectral methods often leave us with implicit answers to the original probabilistic problem, and fractional calculus is applicable only to self-similar Lévy motions. In the present paper, the authors suggest an alternative analytical approach, using the Poisson superposition jump structure of non-Brownian Lévy motions, and which has been preliminary explored by them in some particular examples [J. Stat. Phys. 111, 739--768 (2003; Zbl 1049.82053), Physica A 336, 219 (2004)]. In this paper, they provide a more general and comprehensive view of their proposal and apply it to two exemplary topics: systems governed by Lévy-driven Ornstein-Uhlenbeck dynamics -- systems subject to a restoring field generated by a quadratic potential, and perturbed by a continuous-time random noise -- and systems subject to temporal Lévy subordination -- where the time ticks according to internal subjective clocks rather than according to a universal objective clock --, and which allows for subdiffusive as well as superdiffusive motions. The authors show several advantages of their method -- namely, a combination of spectral analysis and Lévy-Khinchin theorem of jump analysis -- and which allows to obtain answers and insights not attainable with spectral methods alone, or even when these methods are not implementable.