Abstract
In this paper, we consider parallelization of simulated annealing for a product configuration space L S. At every step of the algorithm, each site is activated with probability τ and all the activated sites update synchronously their value. Concerning reversibility conditions and behaviour in the low temperature region, we show that the fully parallel algorithm (τ=1) is a quite singular case and we prove that for the 2-D Ising Model the invariant probability measure at constant temperature T converges to the uniform probability measure on the two global minima when T→0 iff τ<1.
The high temperature region is studied for S=ℤ and translation invariant potential. We construct a sequence of approximations converging towards the invariant probability measure which can be derived from an implementable algorithm and we show that this sequence may be useful to compute valuable approximations of such statistical quantities as the mean energy
Preview
Unable to display preview. Download preview PDF.
References
R. Azencott, Simulated annealing, Séminaire Bourbaki, 40 ième année 697, (1988).
O. Catoni, Large deviations for annealing. Thesis held at university Paris-sud Orsay, March 1990.
D. P. Connors and P. K. Kumar, Simulated annealing-type Markov chains and their order balance equations. SIAM Journal of Control and Optimization, Vol 27, No. 6 (1989) 1440–1461.
R. L. Dobrushin, Prescribing a system of random variables by conditional distributions. Theory of probability and its applications 15 (1970) 458–486.
Pablo A. Ferrari, A. Frigessi and R. H. Schonmann, Convergence of the Partially Parallel Gibbs Sampler with Annealing. (1991). Preprint.
I. Gaudron and A. Trouvé, Parallelization of simulated anncaling for spin glasses energies. In: Simulated annealing, Parallelization techniques, R. Azencott et al. (eds.), John Willey and Sons. To appear in 1992.
S. Geman and D. Geman, Stochastic Relaxation, Gibbs Distribution and the bayesian Restauration of Images. IEEE Proc. Pattern. Anal. Mach. Intell. 6 (1984) 721–741.
B. Gidas, Non stationary Markov chains and convergence of annealing algorithms, Journal of Statistical Physics 39 (1985) 73–131.
B. Hajek, Cooling schedules for optimal annealing. preprints Math. Op. Research (1987)
R. Holley and D. Stroock, Annealing via sobolev Inequalities. Communications in Mathematical Physics 115 (1988)
C.R. Hwang and S.J. Sheu, Singular perturbated Markov chains and exact behaviors of simulated annealing process, preprint, Institute of mathematics, Academia Sinica, Taipei, Taiwan, 1988.
S. Kirkpatrick, C. Gelatt and M. Vecchi, Optimization by simulated annealing, Science 220 (1983) 671–680.
O. Kozlov and N. Vasilyev, Reversible Markov chain with local interaction. In: Multicomponent random systems, R.L. Dobrushin and Ya. G. Sinai eds. (Dekker, New-York) (1980) 451–469.
J.L. Lebowitz, C. Maes and E Speer, Statistical mechanics of probabilistic cellular automata. Journal of Statistical Physics 59 (1990) 117–170.
T.M. Liggett, Interacting Particle System. Berlin, Heidelberg, New York: Springer 1985.
L. Miclo, Evolution de l'énergie libre. Applications à l'étude de la convergence des algorithmes du recuit simulé. Thesis held at university Paris 6, Febuary 1991.
A. Trouvé, Parallel simulated annealing and effective optimization: Study of the limit measure. In: Simulated annealing, Parallelization techniques, R. Azencott et al. (eds.), John Willey and Sons. To appear in 1992.
A. Trouvé, Partially synchronous kernel and parallelization of simulated annealing. Comptes Rendus de l'Académie des Sciences de Paris I 312 (1991) 155–158.
A.D. Wentzell and M.I. Freidlin, Random perturbations of dynamical systems. Springer-Verlag (1984)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1992 Springer-Verlag
About this paper
Cite this paper
Trouvé, A. (1992). Partially parallel simulated annealing: Low and high temperature approach of the invariante measure. In: Karatzas, I., Ocone, D. (eds) Applied Stochastic Analysis. Lecture Notes in Control and Information Sciences, vol 177. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0007063
Download citation
DOI: https://doi.org/10.1007/BFb0007063
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-55296-3
Online ISBN: 978-3-540-47017-5
eBook Packages: Springer Book Archive