Measuring the Irreversibility of Numerical Schemes for Reversible Stochastic Differential Equations

Markos Katsoulakis; Yannis Pantazis; Luc Rey-Bellet

ESAIM: Mathematical Modelling and Numerical Analysis - Modélisation Mathématique et Analyse Numérique (2014)

  • Volume: 48, Issue: 5, page 1351-1379
  • ISSN: 0764-583X

Abstract

top
For a stationary Markov process the detailed balance condition is equivalent to the time-reversibility of the process. For stochastic differential equations (SDE’s), the time discretization of numerical schemes usually destroys the time-reversibility property. Despite an extensive literature on the numerical analysis for SDE’s, their stability properties, strong and/or weak error estimates, large deviations and infinite-time estimates, no quantitative results are known on the lack of reversibility of discrete-time approximation processes. In this paper we provide such quantitative estimates by using the concept of entropy production rate, inspired by ideas from non-equilibrium statistical mechanics. The entropy production rate for a stochastic process is defined as the relative entropy (per unit time) of the path measure of the process with respect to the path measure of the time-reversed process. By construction the entropy production rate is nonnegative and it vanishes if and only if the process is reversible. Crucially, from a numerical point of view, the entropy production rate is an a posteriori quantity, hence it can be computed in the course of a simulation as the ergodic average of a certain functional of the process (the so-called Gallavotti−Cohen (GC) action functional). We compute the entropy production for various numerical schemes such as explicit Euler−Maruyama and explicit Milstein’s for reversible SDEs with additive or multiplicative noise. In addition we analyze the entropy production for the BBK integrator for the Langevin equation. The order (in the time-discretization step Δt) of the entropy production rate provides a tool to classify numerical schemes in terms of their (discretization-induced) irreversibility. Our results show that the type of the noise critically affects the behavior of the entropy production rate. As a striking example of our results we show that the Euler scheme for multiplicative noise is not an adequate scheme from a reversibility point of view since its entropy production rate does not decrease with Δt.

How to cite

top

Katsoulakis, Markos, Pantazis, Yannis, and Rey-Bellet, Luc. "Measuring the Irreversibility of Numerical Schemes for Reversible Stochastic Differential Equations." ESAIM: Mathematical Modelling and Numerical Analysis - Modélisation Mathématique et Analyse Numérique 48.5 (2014): 1351-1379. <http://eudml.org/doc/273314>.

@article{Katsoulakis2014,
abstract = {For a stationary Markov process the detailed balance condition is equivalent to the time-reversibility of the process. For stochastic differential equations (SDE’s), the time discretization of numerical schemes usually destroys the time-reversibility property. Despite an extensive literature on the numerical analysis for SDE’s, their stability properties, strong and/or weak error estimates, large deviations and infinite-time estimates, no quantitative results are known on the lack of reversibility of discrete-time approximation processes. In this paper we provide such quantitative estimates by using the concept of entropy production rate, inspired by ideas from non-equilibrium statistical mechanics. The entropy production rate for a stochastic process is defined as the relative entropy (per unit time) of the path measure of the process with respect to the path measure of the time-reversed process. By construction the entropy production rate is nonnegative and it vanishes if and only if the process is reversible. Crucially, from a numerical point of view, the entropy production rate is an a posteriori quantity, hence it can be computed in the course of a simulation as the ergodic average of a certain functional of the process (the so-called Gallavotti−Cohen (GC) action functional). We compute the entropy production for various numerical schemes such as explicit Euler−Maruyama and explicit Milstein’s for reversible SDEs with additive or multiplicative noise. In addition we analyze the entropy production for the BBK integrator for the Langevin equation. The order (in the time-discretization step Δt) of the entropy production rate provides a tool to classify numerical schemes in terms of their (discretization-induced) irreversibility. Our results show that the type of the noise critically affects the behavior of the entropy production rate. As a striking example of our results we show that the Euler scheme for multiplicative noise is not an adequate scheme from a reversibility point of view since its entropy production rate does not decrease with Δt.},
author = {Katsoulakis, Markos, Pantazis, Yannis, Rey-Bellet, Luc},
journal = {ESAIM: Mathematical Modelling and Numerical Analysis - Modélisation Mathématique et Analyse Numérique},
keywords = {stochastic differential equations; detailed balance; reversibility; relative entropy; entropy production; numerical integration; (overdamped) Langevin process; Euler-Maruyama method; Milstein method},
language = {eng},
number = {5},
pages = {1351-1379},
publisher = {EDP-Sciences},
title = {Measuring the Irreversibility of Numerical Schemes for Reversible Stochastic Differential Equations},
url = {http://eudml.org/doc/273314},
volume = {48},
year = {2014},
}

TY - JOUR
AU - Katsoulakis, Markos
AU - Pantazis, Yannis
AU - Rey-Bellet, Luc
TI - Measuring the Irreversibility of Numerical Schemes for Reversible Stochastic Differential Equations
JO - ESAIM: Mathematical Modelling and Numerical Analysis - Modélisation Mathématique et Analyse Numérique
PY - 2014
PB - EDP-Sciences
VL - 48
IS - 5
SP - 1351
EP - 1379
AB - For a stationary Markov process the detailed balance condition is equivalent to the time-reversibility of the process. For stochastic differential equations (SDE’s), the time discretization of numerical schemes usually destroys the time-reversibility property. Despite an extensive literature on the numerical analysis for SDE’s, their stability properties, strong and/or weak error estimates, large deviations and infinite-time estimates, no quantitative results are known on the lack of reversibility of discrete-time approximation processes. In this paper we provide such quantitative estimates by using the concept of entropy production rate, inspired by ideas from non-equilibrium statistical mechanics. The entropy production rate for a stochastic process is defined as the relative entropy (per unit time) of the path measure of the process with respect to the path measure of the time-reversed process. By construction the entropy production rate is nonnegative and it vanishes if and only if the process is reversible. Crucially, from a numerical point of view, the entropy production rate is an a posteriori quantity, hence it can be computed in the course of a simulation as the ergodic average of a certain functional of the process (the so-called Gallavotti−Cohen (GC) action functional). We compute the entropy production for various numerical schemes such as explicit Euler−Maruyama and explicit Milstein’s for reversible SDEs with additive or multiplicative noise. In addition we analyze the entropy production for the BBK integrator for the Langevin equation. The order (in the time-discretization step Δt) of the entropy production rate provides a tool to classify numerical schemes in terms of their (discretization-induced) irreversibility. Our results show that the type of the noise critically affects the behavior of the entropy production rate. As a striking example of our results we show that the Euler scheme for multiplicative noise is not an adequate scheme from a reversibility point of view since its entropy production rate does not decrease with Δt.
LA - eng
KW - stochastic differential equations; detailed balance; reversibility; relative entropy; entropy production; numerical integration; (overdamped) Langevin process; Euler-Maruyama method; Milstein method
UR - http://eudml.org/doc/273314
ER -

References

top
  1. [1] G. Arampatzis, M.A. Katsoulakis, P. Plechac, M. Taufer and L. Xu, Hierarchical fractional-step approximations and parallel kinetic Monte Carlo algorithms. J. Comput. Phys.231 (2012) 7795–7841. Zbl1259.82110MR2972869
  2. [2] V. Bally and D. Talay, The law of the Euler scheme for stochastic differential equations. I. Convergence rate of the density. Monte Carlo Methods Appl. 2 (1996) 93–128. Zbl0866.60049MR1401964
  3. [3] V. Bally and D. Talay, The law of the Euler scheme for stochastic differential equations. I. Convergence rate of the distribution function. Probab. Theory Related Fields 104 (1996) 43–60. Zbl0838.60051MR1367666
  4. [4] N. Bou-Rabee and E. Vanden-Eijnden, Pathwise accuracy and ergodicity of Metropolized integrators for SDEs. Commut. Pure Appl. Math. LXIII (2010) 0655–0696. Zbl1214.60031MR2583309
  5. [5] A. Brunger, C.B. Brooks and M. Karplus, Stochastic boundary conditions for molecular dynamics simulations of ST2 water. Chem. Phys. Lett.105 (1984) 495–500. 
  6. [6] S. Delong, B.E. Griffith, E. Vanden-Eijnden and A. Donev, Temporal integrators for fluctuating hydrodynamics. Phys. Rev. E 87 (2013) 11. 
  7. [7] G. Gallavotti and E.G.D. Cohen, Dynamical ensembles in nonequilibrium statistical mechanics. Phys. Rev. Lett.74 (1995) 2694–2697. 
  8. [8] C. Gardiner, Handbook of Stochastic Methods: for Physics, Chemistry and the Natural Sciences. Springer Series in Synergetics (1985). Zbl0515.60002MR858704
  9. [9] D.T. Gillespie, Markov Processes: An Introduction for Physical Scientists. Academic Press, New York (1992). Zbl0743.60001MR1133392
  10. [10] E. Hairer, Ch. Lubich and G. Wanner, Structure-preserving algorithms for ordinary differential equations, in Geometric Numerical Integration. vol. 31 of Springer Ser. Comput. Math., 2nd edition. Springer-Verlag, Berlin (2006). Zbl1094.65125MR2221614
  11. [11] V. Jakšić, C.-A. Pillet and L. Rey-Bellet, Entropic fluctuations in statistical mechanics: I. classical dynamical systems. Nonlinearity 2 (2011) 699–763. Zbl1234.37012MR2765481
  12. [12] R. Khasminskii, Stochastic Stability of Differential Equations, 2nd edition. Springer (2010). Zbl1241.60002MR2894052
  13. [13] P.E. Kloeden and E. Platen, Numerical Solution Stochastic Differential Equations, 3rd edition. Springer-Verlag (1999). Zbl0752.60043MR1214374
  14. [14] J.L. Lebowitz and H. Spohn, A Gallavotti-Cohen type symmetry in the large deviation functional for stochastic dynamics. J. Stat. Phys.95 (1999) 333–365. Zbl0934.60090MR1705590
  15. [15] T. Lelievre, M. Rousset and G. Stoltz, Free Energy Computations: A Math. Perspective. Imperial College Press (2010). Zbl1227.82002MR2681239
  16. [16] C. Maes and K. Netočný, Minimum entropy production principle from a dynamical fluctuation law. J. Math. Phys. 48 (2007) 053306. Zbl1144.81384MR2329862
  17. [17] C. Maes, K. Netočný and B. Wynants, Steady state statistics of driven diffusions. Phys. A387 (2008) 2675–2689. MR2587167
  18. [18] C. Maes, F. Redig and A. Van Moffaert, On the definition of entropy production, via examples. J. Math. Phys.41 (2000) 1528–1553. Zbl0977.82025MR1757968
  19. [19] J.C. Mattingly, A.M. Stuart and M.V. Tretyakov, Convergence of numerical time-averaging and stationary measures via Poisson equations. SIAM J. Numer. Anal.48 (2010) 552–577. Zbl1217.65014MR2669996
  20. [20] J.C. Mattingly, A.M. Stuart and D.J. Higham, Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise. Stoch. Process. Appl.101 (2002) 185–232. Zbl1075.60072MR1931266
  21. [21] S.P. Meyn and R.L. Tweedie, Markov Chains and Stochastic Stability. Springer-Verlag (1993). Zbl0925.60001MR1287609
  22. [22] G. Milstein and M. Tretyakov, Stochastic Numerics for Mathematical Physics for Springer (2004). Zbl1085.60004MR2069903
  23. [23] G. Nicolis and I. Prigogine, Self-Organization in Nonequilibrium Systems. Wiley, New York (1977). Zbl0363.93005MR522141
  24. [24] L. Rey-Bellet and L.E. Thomas, Exponential convergence to non-equilibrium stationary states in classical statistical mechanics. Comm. Math. Phys.225 (2002) 305–329. Zbl0989.82023MR1889227
  25. [25] L. Rey-Bellet, Ergodic properties of Markov processes. In Open quantum systems. II, vol. 1881. Lect. Notes Math. Springer, Berlin (2006) 1–39. Zbl1126.60057MR2248986
  26. [26] G.O. Roberts and R.L. Tweedie, Geometric convergence and central limit theorems for multidimensional Hastings and Metropolis algorithms. Biometrika83 (1996) 95–110. Zbl0888.60064MR1399158
  27. [27] T. Schlick, Molecular Modeling and Simulation. Springer (2002). Zbl1011.92019MR1921061
  28. [28] J. Schnakenberg, Network theory of microscopic and macroscopic behavior of master equation systems. Rev. Modern Phys.48 (1976) 571–585. MR443796
  29. [29] Yunsic Shim and J.G. Amar, Semirigorous synchronous relaxation algorithm for parallel kinetic Monte Carlo simulations of thin film growth. Phys. Rev. B71 (2005) 125–432. Zbl1087.82015
  30. [30] D. Talay, Second order discretization schemes of stochastic differential systems for the computation of the invariant law. Stochastics Stochastics Rep.29 (1990) 13–36. Zbl0697.60066
  31. [31] D. Talay, Stochastic Hamiltonian systems: exponential convergence to the invariant measure, and discretization by the implicit Euler scheme. Markov Processes and Related Fields8 (2002) 163–198. Zbl1011.60039MR1924934
  32. [32] D. Talay and L. Tubaro, Expansion of the global error for numerical schemes solving stochastic differential equations. Stoch. Anal. Appl.8 (1990) 483–509. Zbl0718.60058MR1091544
  33. [33] N.G. van Kampen, Stochastic Processes in Physics and Chemistry. North Holland (2006). Zbl0974.60020

NotesEmbed ?

top

You must be logged in to post comments.

To embed these notes on your page include the following JavaScript code on your page where you want the notes to appear.

Only the controls for the widget will be shown in your chosen language. Notes will be shown in their authored language.

Tells the widget how many notes to show per page. You can cycle through additional notes using the next and previous controls.

    
                

                
Note: Best practice suggests putting the JavaScript code just before the closing </body> tag.