Excursions of diffusion processes and continued fractions

From MaRDI portal



Abstract: It is well-known that the excursions of a one-dimensional diffusion process can be studied by considering a certain Riccati equation associated with the process. We show that, in many cases of interest, the Riccati equation can be solved in terms of an infinite continued fraction. We examine the probabilistic significance of the expansion. To illustrate our results, we discuss some examples of diffusions in deterministic and in random environments.


For many diffusions, one can associate with a generator some Riccati equation, the solution to which can be found in terms of certain continued fractions. On the other hand, this problem has a probabilistic counterpart: there is a relationship between the characteristic exponent of a certain Lévy process [\textit{J. Pitman} and \textit{M. Yor}, Bernoulli 9, No. 1, 1--24 (2003; Zbl 1024.60032)] and the solution to such a Riccati equation. Moreover, this solution can be written as a Stieltjes function \(S(w)\) [\textit{R. Schilling} et al., Bernstein functions. Berlin: de Gruyter (2012; Zbl 1257.33001)]. \textit{T. J. Stieltjes} [Toulouse Ann. VIII, J1--J122 (1894); ibid. IX, A1--A47 (1895); C. R. CXVIII, 1401--1403 (1894; JFM 25.0326.01)] considered the problem of recovering a measure on \(\mathbb{R}_+\) from its moments, and to do that used continued fractions of the form \[ \frac{1}{m_0w+ \frac{1}{l_1+ \frac{1}{m_1w + \frac{1}{l_2+\dots}}}}. \] The approach of Stieltjes is used in order to find the solution to the Riccati equation treated in the present paper. Further, some applications are considered, in particular, when the initial generator is that of a diffusion with drift, or of a Bessel process.



Cites work









This page was built for publication: Excursions of diffusion processes and continued fractions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q720745)