1: \begin{abstract}

2: %DECOR

3: Experimental studies and numerical simulations have been devoted to the motion of an object subjected to a dry friction and an external random force. The experimental and numerical observations suggest that the variance of the object displacement grows linearly with time. Here, the variance growth rate is called diffusivity. 

4: %PROBLEM/OBJECTIF

5: The goal of this paper is to propose efficient stochastic simulation methods for computing the diffusivity when the external random force is white or colored and to provide a theoretical justification supporting experimental observations.

6: %CADRE MATHEMATIQUE

7: %We use the framework of piecewise deterministic Markov process (PDMP) to model dry friction as recently proposed in [].

8: %The forcing is a colored noise represented by a pure jump process which is itself a discrete (in space) approximation of an Ornstein-Uhlenbeck (OU) noise. 

9: %RESULTS

10: We show that the variance of the object displacement has a linear growth in time.

11: We obtain a representation formula for the diffusivity as an expectation of a functional of an excursion. This makes it possible to design a consistent estimator of the diffusivity with an associated confidence interval. 

12: We show that the estimator has much better convergence properties than usual Monte Carlo estimators.

13: From highly accurate simulations we can then propose and discuss scaling relations between the diffusivity, noise strength and correlation time.

14: %As a byproduct, we derive an effective ad-hoc Monte Carlo approach for its computation. 

15: %Then, the notion of long excursion and the resulting numerical approach are extrapolated from the PDMP to the limiting OU noise driven system and studied via numerical experiments. The relation between the noise strength and the diffusivity is studied as well. 

16: \end{abstract}

17: