Article Contents
Article Contents

# Exit problem for Ornstein-Uhlenbeck processes: A random walk approach

• * Corresponding author: Samuel Herrmann
• In order to approximate the exit time of a one-dimensional diffusion process, we propose an algorithm based on a random walk. Such an algorithm so-called Walk on Moving Spheres was already introduced in the Brownian context. The aim is therefore to generalize this numerical approach to the Ornstein-Uhlenbeck process and to describe the efficiency of the method.

Mathematics Subject Classification: Primary: 65C05; Secondary: 60J60, 60G40, 60G46.

 Citation:

• Figure 1.  A sample of the algorithm for the O.U. exit time with parameters $\theta = 0.1$ and $\sigma = 1$. We observe the walk on spheres associated with the diffusion process starting at $x = 5$ and moving in the interval $[2,7]$. The algorithm corresponding to $\epsilon = 0,5$ is represented by the plain style spheroids whereas the case $\epsilon = 10^{-3}$ corresponds to the whole sequence of spheroids. In both cases we set $\gamma = 10^{-6}$

Figure 2.  Histogram of the outcome variable for the O.U. with parameters $\theta = 0.1$ and $\sigma = 1$ when the stopped diffusion process starts at 5 and involves in the interval [2, 7] with $\epsilon = 10^{-3}$ and $\gamma = 10^{-6}$

Figure 3.  Histogram of the approximated first exit time of the interval $[a,b]$ using the WOMS algorithm and approximated p.d.f. of the first passage time through the level $b$ (curve). Here $X_0 = -3$, $\theta = 1$, $\sigma = 1$ and $[a,b] = [-10,-1]$

Figure 4.  Simulation of the O.U. exit time from the interval $[2,7]$. The starting position is $X_0 = 5$ and the parameters are given by $\theta = 0.1$, $\sigma = 1$ and $\gamma = 10^{-6}$. Histogram of the number of steps observed for $\epsilon = 10^{-3}$

Figure 5.  Simulation of the O.U. exit time from the interval $[2,7]$. The starting position is $X_0 = 5$ and the parameters are given by $\theta = 0.1$, $\sigma = 1$ and $\gamma = 10^{-6}$. Average number of steps versus $\epsilon$ (in logarithmic scale)

Figure 6.  Error bound $\Xi$ versus $\epsilon$ for different values of $\theta$ with $\sigma = 1$, $a = -1$, $b = 1$, $\gamma = 1$

•  [1] P. Baldi, L. Caramellino and M. G. Iovino, Pricing general barrier options: A numerical approach using sharp large deviations, Math. Finance, 9 (1999), 293-322.  doi: 10.1111/1467-9965.t01-1-00071. [2] I. Binder and M. Braverman, The rate of convergence of the walk on spheres algorithm, Geom. Funct. Anal., 22 (2012), 558-587.  doi: 10.1007/s00039-012-0161-z. [3] A. Buonocore, A. G. Nobile and L. M. Ricciardi, A new integral equation for the evaluation of first-passage-time probability densities, Adv. in Appl. Probab., 19 (1987), 784-800.  doi: 10.2307/1427102. [4] M. Deaconu and S. Herrmann, Simulation of hitting times for bessel processes with non-integer dimension, Bernoulli, 23 (2017), 3744-3771.  doi: 10.3150/16-BEJ866. [5] S. Herrmann and C. Zucca, Exact simulation of the first-passage time of diffusions, J. Sci. Comput., 79 (2019), 1477-1504.  doi: 10.1007/s10915-018-00900-3. [6] I. Karatzas and S. E. Shreve, Brownian Motion and Stochastic Calculus, volume 113 of Graduate Texts in Mathematics, second edition, Springer-Verlag, New York, 1991. doi: 10.1007/978-1-4612-0949-2. [7] H. R. Lerche, Boundary Crossing of Brownian Motion, volume 40 of Lecture Notes in Statistics, Springer-Verlag, Berlin, 1986. doi: 10.1007/978-1-4615-6569-7. [8] M. Motoo, Some evaluations for continuous Monte Carlo method by using Brownian hitting process, Ann. Inst. Statist. Math. Tokyo, 11 (1959), 49-54.  doi: 10.1007/BF01831723. [9] M. E. Muller, Some continuous Monte Carlo methods for the Dirichlet problem, Ann. Math. Statist., 27 (1956), 569-589.  doi: 10.1214/aoms/1177728169. [10] J. R. Norris,  Markov Chains, Cambridge University Press, Cambridge, 1998. [11] D. Revuz and M. Yor, Continuous Martingales and Brownian Motion, volume 293 of Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], Springer-Verlag, Berlin, third edition, 1999. doi: 10.1007/978-3-662-06400-9. [12] K. K. Sabelfeld and N. A. Simonov, Random Walks on Boundary for Solving PDEs, VSP, Utrecht, 1994. [13] K. K. Sabelfeld, Monte Carlo Methods in Boundary Value Problems, Springer Series in Computational Physics. Springer-Verlag, Berlin, 1991. [14] L. Sacerdote, O. Telve and C. Zucca, Joint densities of first hitting times of a diffusion process through two time-dependent boundaries, Adv. in Appl. Probab., 46 (2014), 186-202.  doi: 10.1239/aap/1396360109.

Figures(6)