Simulated annealing algorithms and Markov chains with rare transitions

In these notes, written for a D.E.A. course at University Paris XI during the first term of 1995, we prove the essentials about stochastic optimisation algorithms based on Markov chains with rare transitions, under the weak assumption that the transition matrix obeys a large deviation principle. We present a new simplified line of proofs based on the Freidlin and Wentzell graphical approach. The case of Markov chains with a periodic behaviour at null temperature is considered. We have also included some pages about the spectral gap approach where we follow Diaconis and Stroock [13] and Ingrassia [23] in a more conventional way, except for the application to non reversible Metropolis algorithms (subsection 6.2.2) where we present an original result.

[1]  G. Kirchhoff.  Ueber die Auflösung der Gleichungen, auf welche man bei der Untersuchung der linearen Vertheilung galvanischer Ströme geführt wird , 1847 .

[2]  C. D. Gelatt,et al.  Optimization by Simulated Annealing , 1983, Science.

[3]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  M. Freidlin,et al.  Random Perturbations of Dynamical Systems , 1984 .

[5]  D. Stroock,et al.  Simulated annealing via Sobolev inequalities , 1988 .

[6]  D. Stroock,et al.  Asymptotics of the spectral gap with applications to the theory of simulated annealing , 1989 .

[7]  P. Diaconis,et al.  Geometric Bounds for Eigenvalues of Markov Chains , 1991 .

[8]  J. A. Fill.  Eigenvalue bounds on convergence to stationarity for nonreversible markov chains , 1991 .

[9]  L. Miclo.  Evolution de l'énergie libre, applications à l'étude de la convergence des algorithmes du recuit simulé , 1991 .

[10]  C. Hwang,et al.  Singular perturbed Markov chains and exact behaviors of simulated annealing processes , 1992 .

[11]  O. Catoni.  Rough Large Deviation Estimates for Simulated Annealing: Application to Exponential Schedules , 1992 .

[12]  O. Catoni,et al.  Exponential triangular cooling schedules for simulated annealing algorithms : A case study , 1992 .

[13]  A. Trouvé.  Parallélisation massive du recuit simulé , 1993 .

[14]  A. Trouvé.  Rough Large Deviation Estimates for the Optimal Convergence Speed Exponent of Generalized Simulated , 1994 .

[15]  S. Ingrassia.  ON THE RATE OF CONVERGENCE OF THE METROPOLIS ALGORITHM AND GIBBS SAMPLER BY GEOMETRIC BOUNDS , 1994 .

[16]  J. Deuschel,et al.  $L^2$ Convergence of Time Nonhomogeneous Markov Processes: I. Spectral Estimates , 1994 .

[17]  Olivier Catoni,et al.  Metropolis, Simulated Annealing, and Iterated Energy Transformation Algorithms: Theory and Experiments , 1996, J. Complex..

[18]  A. Trouvé.  Cycle Decompositions and Simulated Annealing , 1996 .

[19]  L. Miclo.  Sur les problèmes de sortie discrets inhomogènes , 1996 .

[20]  L. Miclo,et al.  SUR LES PROBL EMES DE SORTIE DISCRETS INHOMOG ENES , 1996 .

[21]  L. Miclo,et al.  Remarques sur l’hypercontractivité et l’évolution de l’entropie pour des chaînes de Markov finies , 1997 .

[22]  O. Catoni,et al.  The exit path of a Markov chain with rare transitions , 1997 .

[23]  L. Saloff-Coste,et al.  Lectures on finite Markov chains , 1997 .

[24]  O. Catoni,et al.  Piecewise constant triangular cooling schedules for generalized simulated annealing algorithms , 1998 .

[25]  O. Catoni.  The energy transformation method for the Metropolis algorithm compared with Simulated Annealing , 1998 .

[26]  M. Laurent.  Sur les temps d'occupations des processus de markov finis inhomogènes à basse température , 1998 .

[27]  O. Catoni.  Solving Scheduling Problems by Simulated Annealing , 1998 .