Abstract
Combining classical arguments for the analysis of the simulated annealing algorithm with the more recent hypocoercive method of distorted entropy, we prove the convergence for large time of the kinetic Langevin annealing with logarithmic cooling schedule.
Similar content being viewed by others
Notes
Allowing large steps is not a reasonable solution, since in applied problems the dimension is large and the reasonable configurations (i.e. the points where U is not too large) lie in a very small area in view of the Lebesgue measure. A uniformly-generated jump proposal will always be absurd, and rejected.
References
Ané, C., Blachère, S., Chafaï, D., Fougères, P., Gentil, I., Malrieu, F., Roberto, C., Scheffer, G.: Sur les inégalités de Sobolev logarithmiques. Panoramas et synthèses. Société mathématique de France, Paris. http://opac.inria.fr/record=b1097031 (2000)
Arnold, A., Erb, J.: Sharp entropy decay for hypocoercive and non-symmetric Fokker–Planck equations with linear drift. ArXiv e-prints (2014)
Bakry, D., Gentil, I., Ledoux, M.: Analysis and geometry of Markov diffusion operators. In: Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], vol. 348. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-00227-9
Baudoin, F.: Bakry–Emery meet Villani. ArXiv e-prints (2013)
Bolley, F., Gentil, I.: Phi-entropy inequalities for diffusion semigroups. J. Math. Pures Appl. (9) 93(5), 449–473 (2010). https://doi.org/10.1016/j.matpur.2010.02.004
Catoni, O.: Rough large deviation estimates for simulated annealing: application to exponential schedules. Ann. Probab. 20(3), 1109–1146 (1992). http://links.jstor.org/sici?sici=0091-1798(199207)20:3<1109:RLDEFS>2.0.CO;2-Z&origin=MSN
Chiang, T.S., Hwang, C.R., Sheu, S.: Diffusion for global optimization in \({\mathbb{R}}^n\). SIAM J. Control Optim. 25(3), 737–753 (1987). https://doi.org/10.1137/0325042
Dolbeault, J., Mouhot, C., Schmeiser, C.: Hypocoercivity for kinetic equations with linear relaxation terms. C. R. Math. Acad. Sci. Paris 347(9–10), 511–516 (2009). https://doi.org/10.1016/j.crma.2009.02.025
Fort, G., Jourdain, B., Kuhn, E., Lelièvre, T., Stoltz, G.: Efficiency of the Wang–Landau algorithm: a simple test case. AMRX 2, 275–311 (2014)
Gadat, S., Miclo, L.: Spectral decompositions and \(\mathbb{L}^2\)-operator norms of toy hypocoercive semi-groups. Kinet. Relat. Models 6(2), 317–372 (2013)
Gadat, S., Panloup, F.: Long time behaviour and stationary regime of memory gradient diffusions. Ann. Inst. Henri Poincaré Probab. Stat. 50(2), 564–601 (2014). https://doi.org/10.1214/12-AIHP536
Guillin, A., Monmarché, P.: Optimal linear drift for an hypoelliptic diffusion. Electron. Commun. Probab. 21, 1–14 (2016)
Holley, R.A., Kusuoka, S., Stroock, D.W.: Asymptotics of the spectral gap with applications to the theory of simulated annealing. J. Funct. Anal. 83(2), 333–347 (1989). https://doi.org/10.1016/0022-1236(89)90023-2
Holley, R.A., Stroock, D.W.: Simulated annealing via Sobolev inequalities. Commun. Math. Phys. 115(4), 553–569 (1988). http://projecteuclid.org/euclid.cmp/1104161084
Kuwada, K.: Duality on gradient estimates and Wasserstein controls. J. Funct. Anal. 258(11), 3758–3774 (2010). https://doi.org/10.1016/j.jfa.2010.01.010
Lelièvre, T., Rousset, M., Stoltz, G.: Long-time convergence of an adaptive biasing force method. Nonlinearity 21(6), 1155–1181 (2008). https://doi.org/10.1088/0951-7715/21/6/001
Menz, G., Schlichting, A.: Poincaré and logarithmic Sobolev inequalities by decomposition of the energy landscape. Ann. Probab. 42(5), 1809–1884 (2014). https://doi.org/10.1214/14-AOP908
Miclo, L.: Recuit simulé sur \(R^n\). Étude de l’évolution de l’énergie libre. Ann. Inst. H. Poincaré Probab. Statist. 28(2), 235–266 (1992). http://www.numdam.org/item?id=AIHPB_1992__28_2_235_0
Monmarché, P.: Generalized \(\Gamma \) calculus and application to interacting particles on a graph. ArXiv e-prints (2015)
Monmarché, P.: Piecewise deterministic simulated annealing. ALEA Lat. Am. J. Probab. Math. Stat. 13(1), 357–398 (2016)
Robbe, V.: Small eigenvalues of the low temperature linear relaxation boltzmann equation with a confining potential. Annales Henri Poincaré 17(4), 937–952 (2016). https://doi.org/10.1007/s00023-015-0410-4
Scemama, A., Lelièvre, T., Stoltz, G., Caffarel, M.: An efficient sampling algorithm for variational Monte Carlo. J. Chem. Phys. 125, 114105 (2006)
Sun, Y., Garcia, A.: Interactive diffusions for global optimization. J. Optim. Theory Appl. 163(2), 491–509 (2014). https://doi.org/10.1007/s10957-013-0394-5
Talay, D.: Stochastic Hamiltonian systems: exponential convergence to the invariant measure, and discretization by the implicit Euler scheme. Markov Process. Relat. Fields 8(2), 163–198 (2002)
Taniguchi, S.: Applications of Malliavin’s calculus to time-dependent systems of heat equations. Osaka J. Math. 22(2), 307–320 (1985). http://projecteuclid.org/euclid.ojm/1200778261
Villani, C.: Hypocoercivity, vol. 202, no. 950. Memoirs of the American Mathematical Society, Providence (2009). https://doi.org/10.1090/S0065-9266-09-00567-5
Wu, L.: Large and moderate deviations and exponential convergence for stochastic damping Hamiltonian systems. Stoch. Process. Appl. 91(2), 205–238 (2001). https://doi.org/10.1016/S0304-4149(00)00061-2
Zitt, P.A.: Annealing diffusions in a potential function with a slow growth. Stoch. Process. Appl. 118(1), 76–119 (2008). https://doi.org/10.1016/j.spa.2007.04.002
Acknowledgements
The author would like to thank Laurent Miclo, who initiated this work, for numerous fruitful discussions. This work has been supported by EFI Project ANR-17-CE40-0030 of the French National Research Agency (ANR).
Author information
Authors and Affiliations
Corresponding author
Additional information
This work has been supported by ANR STAB.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Monmarché, P. Hypocoercivity in metastable settings and kinetic simulated annealing. Probab. Theory Relat. Fields 172, 1215–1248 (2018). https://doi.org/10.1007/s00440-018-0828-y
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00440-018-0828-y