Dans cet article, nous étudions les mesures homogènes de Gibbs sur un arbre de Cayley soumises à une évolution de Glauber à une température infinie, et nous considérons leurs propriétés dites « non Gibbsiennes ». Nous montrons que l' état de Gibbs intermédiaire (c'est à dire pour un champ magnétique nul l'état de Gibbs correspondant à la condition au bord libre) se comporte différemment des états de Gibbs « plus » et « moins ». Par exemple, lorsque le temps est assez grand, toutes les configurations sont mauvaises pour l'état intermédiaire, tandis que la configuration « plus » n'est jamais mauvaise pour l'état « plus ». De plus nous montrons que, pour chaque état, il y a deux transitions. Pour l'état intermédiaire il y a une première transition d'un régime Gibbsien à un régime non-Gibbsien, où certaines configurations mais pas toutes sont mauvaises. Après cette première transition, il y en a une seconde dans laquelle l'état intermédiaire passe à un régime où toutes les configurations sont mauvaises. Pour les états « plus » et « moins », il y a également deux transitions : une première d'un régime Gibbsien à un régime non-Gibbsien, et une deuxième d'un régime non-Gibbsien à un régime Gibbsien.
In this paper we study homogeneous Gibbs measures on a Cayley tree, subjected to an infinite-temperature Glauber evolution, and consider their (non-)Gibbsian properties. We show that the intermediate Gibbs state (which in zero field is the free-boundary-condition Gibbs state) behaves differently from the plus and the minus state. E.g. at large times, all configurations are bad for the intermediate state, whereas the plus configuration never is bad for the plus state. Moreover, we show that for each state there are two transitions. For the intermediate state there is a transition from a Gibbsian regime to a non-Gibbsian regime where some, but not all configurations are bad, and a second one to a regime where all configurations are bad. For the plus and minus state, the two transitions are from a Gibbsian regime to a non-Gibbsian one and then back to a Gibbsian regime again.
@article{AIHPB_2012__48_3_774_0, author = {van Enter, Aernout C. D. and Ermolaev, Victor N. and Iacobelli, Giulio and K\"ulske, Christof}, title = {Gibbs-non-Gibbs properties for evolving Ising models on trees}, journal = {Annales de l'I.H.P. Probabilit\'es et statistiques}, volume = {48}, year = {2012}, pages = {774-791}, doi = {10.1214/11-AIHP421}, mrnumber = {2976563}, zbl = {1255.82037}, language = {en}, url = {http://dml.mathdoc.fr/item/AIHPB_2012__48_3_774_0} }
van Enter, Aernout C. D.; Ermolaev, Victor N.; Iacobelli, Giulio; Külske, Christof. Gibbs-non-Gibbs properties for evolving Ising models on trees. Annales de l'I.H.P. Probabilités et statistiques, Tome 48 (2012) pp. 774-791. doi : 10.1214/11-AIHP421. http://gdmltest.u-ga.fr/item/AIHPB_2012__48_3_774_0/
[1] On the purity of the limiting Gibbs state for the Ising model on the Bethe lattice. J. Statist. Phys. 79 (1995) 473-482. | MR 1325591 | Zbl 1081.82515
, and .[2] Propagation of Gibbsianness for infinite-dimensional gradient Brownian diffusions. J. Statist. Phys. 121 (2005) 511-551. | MR 2185338 | Zbl 1127.82039
and .[3] Low-temperature dynamics of the Curie-Weiss model: Periodic orbits, multiple histories and loss of Gibbsianness. J. Statist. Phys. 141 (2010) 727-756. | MR 2739308 | Zbl 1208.82042
and .[4] Gibbsianness and non-Gibbsianness in lattice random fields. In Les Houches Summer School, Session LXXXIII, 2005. Mathematical Statistical Physics, A. Elsevier, Amsterdam, 2006. | MR 2581896 | Zbl pre05723808
.[5] Gibbs Measures and Phase Transitions. de Gruyter, Berlin, 1988. ISBN 0-89925-462-4. | MR 956646 | Zbl 1225.60001
.[6] Almost sure quasilocality fails for the random-cluster model on a tree. J. Statist. Phys. 84 (1996) 1351-1361. | MR 1412082 | Zbl 1081.82520
.[7] Gibbs properties of the fuzzy Potts model on trees and in mean field. Markov Process. Related Fields 10 (2004) 477-506. | MR 2097868 | Zbl 1210.82023
and .[8] On the extremality of the disordered state for the Ising model on the Bethe lattice. Lett. Math. Phys. 37 (1996) 137-143. | MR 1391195 | Zbl 0848.60090
.[9] The posterior metric and the goodness of Gibbsianness for transforms of Gibbs measures. Electron. J. Probab. 13 (2008) 1307-1344. | MR 2438808 | Zbl 1190.60096
and .[10] Loss without recovery of Gibbsianness during diffusion of continuous spins. Probab. Theory Related Fields 135 (2006) 428-456. | MR 2240694 | Zbl 1095.60027
and .[11] Fractal failure of quasilocality for a majority rule transformation on a tree. Lett. Math. Phys. 54 (2000) 11-24. | MR 1846719 | Zbl 0989.82013
.[12] Short time conservation of Gibbsianness under local stochastic evolutions. J. Statist. Phys. 109 (2002) 1073-1090. | MR 1938286 | Zbl 1015.60097
and .[13] On Gibbs measures of transforms of lattice and mean-field systems. Ph.D. thesis, Rijksuniversiteit Groningen, 2009. | Zbl 1192.82004
.[14] Robust phase tramsitions for Heisenberg and other models on general trees. Ann. Probab. 27 (1999) 876-912. | MR 1698979 | Zbl 0981.60096
and .[15] Short-time Gibbsianness for infinite-dimensional diffusions with space-time interaction. J. Statist. Phys. 138 (2010) 1124-1144. | MR 2601426 | Zbl 1188.82047
, and .[16] Loss and recovery of Gibbsianness for spins in small external fields. J. Math. Phys. 49 (2008) 125208. | MR 2484339 | Zbl 1159.81345
and .[17] Gibbsianness versus non-Gibbsianness of time-evolved planar rotor models. Stochastic Processes Appl. 119 (2009) 1866-1888. | MR 2519348 | Zbl 1173.82007
and .[18] Possible loss and recovery of Gibbsianness during the stochastic evolution of Gibbs measures. Comm. Math. Phys. 226 (2002) 101-130. | MR 1889994 | Zbl 0990.82018
, , and .[19] A large-deviation view on dynamical Gibbs-non-Gibbs transitions. Mosc. Math. J. 10 (2010) 687-711. | MR 2791053 | Zbl 1221.82046
, , and .[20] Regularity properties and pathologies of position-space renormalization-group transformations: Scope and limitations of Gibbsian theory. J. Statist. Phys. 72 (1993) 879-1167. | MR 1241537 | Zbl 1101.82314
, and .[21] Gibbs-non-Gibbs properties for -vector lattice and mean-field models. Braz. J. Probab. Stat. 24 (2010) 226-255. | MR 2643565 | Zbl 1200.82015
, , and .