plaquette models for glasses and constraint satisfaction
play

PLAQUETTE MODELS FOR GLASSES AND CONSTRAINT - PowerPoint PPT Presentation

PLAQUETTE MODELS FOR GLASSES AND CONSTRAINT SATISFACTION PROBLEMS Giacomo Gradenigo With: Silvio Franz (LPTMS,Orsay) Stefano Spigler (LPTMS,Orsay).


  1. PLAQUETTE ¡MODELS ¡FOR ¡GLASSES ¡ ¡ AND ¡ ¡ CONSTRAINT ¡SATISFACTION ¡PROBLEMS ¡ Giacomo ¡Gradenigo ¡ ¡ ¡ With: ¡Silvio ¡Franz ¡(LPTMS,Orsay) ¡ Stefano ¡Spigler ¡(LPTMS,Orsay). ¡ LIPhy, ¡27-­‑01-­‑2015 ¡

  2. WHAT ¡IS ¡A ¡GLASS ¡? ¡ log( η ) T g /T …a new phase of matter !!! …Just a very viscous fluid !!! Thermodynamic ¡theories ¡ ¡ KineVcally ¡Constrained ¡Models ¡ (Replicas ¡and ¡all ¡of ¡that…) ¡

  3. Thermodynamic ¡theories ¡ ¡ KineVcally ¡ (Replicas ¡and ¡all ¡of ¡that…) ¡ Constrained ¡Models ¡ ¡ Ÿ Well ¡defined ¡Hamiltonian ¡ Ÿ Set ¡of ¡microscropic ¡ and ¡parVVon ¡funcVon ¡ dynamic ¡rules ¡ Ÿ Ideal ¡glass ¡transiVon, ¡ landscape ¡scenario, ¡predicVons ¡ Ÿ Glassy ¡behaviour ¡at ¡low ¡ on ¡the ¡dynamics ¡ temperatures ¡but ¡no ¡glass-­‑ transiVon ¡ Ÿ Exactly ¡solved ¡only ¡in ¡ approximate ¡theories ¡ ¡ Ÿ Exactly ¡solved ¡in ¡finite ¡ (mean-­‑field, ¡random ¡graphs) ¡ space ¡dimensions ¡ Recent ¡result ¡on ¡jamming ¡(hard ¡spheres ¡in ¡d=∞): ¡ ¡ Review: ¡J.P. ¡Garrahan, ¡P. ¡Sollich, ¡C. ¡ P. ¡Charbonneau, ¡ ¡J.Kurchan, ¡G.Parisi, ¡P.Urbani, ¡ Toninelli, ¡arXiv:1009.6113 ¡ F.Zamponi, ¡Nature ¡Comm. ¡(2014): ¡arXiv.1404.6809 ¡

  4. Thermodynamic ¡theories ¡ ¡ KineVcally ¡ (Replicas ¡and ¡all ¡of ¡that…) ¡ Constrained ¡Models ¡ ¡ Ÿ Well ¡defined ¡Hamiltonian ¡ Ÿ Set ¡of ¡microscropic ¡ and ¡parVVon ¡funcVon ¡ dynamic ¡rules ¡ Ÿ Ideal ¡glass ¡transiVon, ¡ landscape ¡scenario, ¡predicVons ¡ Triangular Plaquette Model Ÿ Glassy ¡ on ¡the ¡dynamics ¡ behaviour ¡at ¡low ¡ temperatures ¡but ¡ no ¡glass-­‑transiVon ¡ Ÿ Exactly ¡solved ¡only ¡in ¡ approximate ¡theories ¡ ¡ Ÿ Exactly ¡solved ¡in ¡finite ¡ (mean-­‑field, ¡random ¡graphs) ¡ space ¡dimensions ¡ The two scenarios are complementary rather than alternative

  5. TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ M. ¡Newman ¡and ¡C. ¡Moore, ¡Phys. ¡Rev. ¡E ¡60, ¡5068 ¡ ¡(1999) ¡ J.P. ¡Garrahan ¡and ¡M. ¡Newman, ¡Phys. ¡Rev. ¡E ¡62, ¡7670 ¡(2000) ¡ X σ j H = − σ i σ j σ k h ijk i σ i = ± 1 σ i σ k WHY ¡INTERSTING? ¡ No ¡phase ¡transiVons ¡ ε p = − σ i ( p ) σ j ( p ) σ k ( p ) 2D: ¡Exactly ¡solvable ¡ X e − β Σ p ε p Z = { ε } A/T 2 � � GLASSY ¡DYNAMICS ¡ τ rel ∼ exp

  6. WHY ¡RELAXATION ¡IS ¡SLOW? ¡ ε p = − σ i ( p ) σ j ( p ) σ k ( p ) (ACTIVATED ¡TRANSPORT ¡OF ¡DEFECTS) ¡ ✚ ¡ ε = 1 Spin ¡to ¡be ¡flipped ¡ (2) ¡ (1) ¡ ✚ ¡ ✚ ¡ ✚ ¡ ✚ ¡ ∆ ε = 2 ∆ ε = 3 ✚ ¡ (4) ¡ (3) ¡ ∆ ε = 1 ∆ ε = 2 ✚ ¡ ✚ ¡ ✚ ¡

  7. ε p = − σ i ( p ) σ j ( p ) σ k ( p ) Effective Kinetic Constraint : creation of a defect is favoured only when there are ✚ ¡ ε = 1 already two closeby Spin ¡to ¡be ¡flipped ¡ (2) ¡ (1) ¡ ✚ ¡ ✚ ¡ ✚ ¡ ✚ ¡ ∆ ε = 2 ∆ ε = 3 ✚ ¡ τ rel ∼ e β ∆ E ( β ) (4) ¡ ∆ ε = 1 ∆ E ( β ) ∼ log[ ξ ( β )] c def ( β ) ∼ e − β = ⇒ ξ ( β ) ∼ e β ✚ ¡ τ rel = e β log ξ = [ ξ ( β )] β = exp( β 2 )

  8. TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ X σ j H = − σ i σ j σ k h ijk i σ k σ i A ¡RECENT ¡RESULT ¡ N � ( a ) � ( b ) H TOT = H [ � ( a ) ] + H [ � ( b ) ] − ✏ X i i 2 i =1 self-dual line 1.5 Two ¡‘’replicas’’ ¡are ¡coupled ¡ low overlap critical T 1 Arbitrary ¡small ¡coupling ¡leads ¡to ¡ endpoint correlaVon ¡between ¡the ¡two ¡replicas ¡ high 0.5 overlap J.P. ¡Garrahan, ¡Phys. ¡Rev. ¡E ¡89, ¡030301(R) ¡(2014) ¡ 0 0 0.5 1 1.5 2 !

  9. TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ X σ j H = − σ i σ j σ k h ijk i σ k σ i OUR ¡PURPOSE ¡ X X H = H 0 + H 1 = − J 0 σ i σ j σ k − J 1 σ i ( r ) σ j ( r ) σ k ( r ) r h ijk i H 1 Show ¡that ¡with ¡ arbitrarily ¡small ¡perturbaVons ¡the ¡ ⌧ 1 system ¡has ¡a ¡finite-­‑temperature ¡glass ¡transiVon. ¡ H 0 Marginal ¡nature ¡of ¡the ¡model. ¡

  10. THE ¡FORMATION ¡OF ¡A ¡GLASS ¡PHASE ¡ (configuraVonal) ¡ENTROPY ¡CRISIS ¡ Energy ¡landscape ¡with ¡‘’many’’ ¡local ¡minima ¡ N = exp [ N Σ ( T )] f = U − T ( s + Σ ) s # valleys = Σ Entropic ¡drive ¡vs ¡Energy ¡Barriers: ¡ ConfiguraVonal ¡ Σ ( T ) > 0 system ¡remains ¡ergodic ¡ Entropy ¡ System ¡becomes ¡a ¡glass ¡ Σ ( T ) = 0 IDEAL ¡GLASS ¡TRANSITION ¡

  11. THE ¡FORMATION ¡OF ¡A ¡GLASS ¡PHASE ¡ ExtrapolaVon ¡of ¡high-­‑temperature ¡thermodynamic ¡potenVals ¡ ConfiguraVonal ¡entropy ¡crisis ¡at ¡a ¡finite ¡temperature: ¡glass ¡formaVon ¡ Σ ( T ) = S liq ( T ) − S sol ( T )

  12. RANDOM-­‑DILUTED ¡ ¡ TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ In ¡the ¡new ¡triplets ¡spins ¡are ¡choosen ¡randomly ¡across ¡the ¡system ¡ N s ≤ N M X X H = H 0 + H 1 = − σ i ( s ) σ j ( s ) σ k ( s ) − σ i ( r ) σ j ( r ) σ k ( r ) s =1 r =1 A ¡diluVon ¡of ¡the ¡plaquenes ¡in ¡the ¡2D ¡laoce ¡is ¡allowed ¡ N s < N

  13. RANDOM-­‑DILUTED ¡ ¡ TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ In ¡the ¡new ¡triplets ¡spins ¡are ¡choosen ¡randomly ¡across ¡the ¡system ¡ N M X X H = − σ i ( s ) σ j ( s ) σ k ( s ) − σ i ( r ) σ j ( r ) σ k ( r ) s =1 r =1 2 3 Z h.T. = 2 N [cosh( β )] M X X h ε p i + tanh 2 ( β ) 4 1 � tanh( β ) h ε p ε q i + . . . 5 p =1 p 6 = q ε p = − σ i ( p ) σ j ( p ) σ k ( p ) Z h.T. = 2 N [cosh( β )] M PROVED ¡! ¡ Entropy ¡crisis ¡at ¡ ✓ ◆ 1 − M S liq > Σ S liq ( β → ∞ ) = finite ¡temperature: ¡ N Glass ¡transiBon? ¡

  14. RANDOM-­‑DILUTED ¡ ¡ TRIANGULAR ¡PLAQUETTE ¡MODEL ¡ exp( A/T 2 ) vs exp( A/ ( T − T K )) = ? -0.7 cooling heating high temperature expansion L = 64 -0.75 L = 256 -0.8 -0.85 E/N spin � rel -0.9 -0.95 0.1 -1 -1.05 1/T 2 -1.1 0 0.2 0.4 0.6 0.8 1 1.2 T m T Z h.T. = 2 N [cosh( β )] M PROVEN ¡ Entropy ¡crisis ¡at ¡ ✓ ◆ 1 − M S liq > Σ S liq ( β → ∞ ) = finite ¡temperature: ¡ N Glass ¡transiBon? ¡

  15. HOW ¡TO ¡HANDLE ¡THE ¡PROBLEM? ¡ TOOLS ¡AND ¡IDEAS ¡FROM ¡ ¡ CONSTRAINT ¡SATISFACTION ¡PROBLEMS! ¡ ¡

  16. DILUTED ¡P-­‑SPIN ¡ FuncVon ¡nodes: ¡Plaquenes ¡ Variable ¡nodes: ¡Spins ¡ BiparVte ¡graph ¡ α = M/N M ¡= ¡number ¡of ¡plaquenes ¡ N ¡= ¡number ¡of ¡spins ¡ X H = − J ijk σ i σ j σ k ijk Each ¡plaquene ¡is ¡anached ¡to ¡ three ¡spins ¡ p ( ` ) = e − 3 ↵ (3 ↵ ) ` Spin ¡connecVvity ¡= ¡Poissonian ¡ ` ! Supercooled ¡ Liquid ¡ liquid ¡ Glass ¡ T d T K

  17. DILUTED ¡P-­‑SPIN ¡ log(N) ¡loops ¡ BiparVte ¡graph ¡ α = M/N X H = − J ijk σ i σ j σ k ijk Each ¡plaquene ¡is ¡anached ¡to ¡ three ¡spins ¡ p ( ` ) = e − 3 ↵ (3 ↵ ) ` Spin ¡connecVvity ¡= ¡Poissonian ¡ ` ! Supercooled ¡ Liquid ¡ liquid ¡ Glass ¡ T d T K

  18. XORSAT ¡ log(N) ¡loops ¡ BiparVte ¡graph ¡ α = M/N System ¡of ¡M ¡linear ¡equaVons ¡ with ¡boolean ¡variables ¡ M. ¡Mezard, ¡F. ¡Ricci-­‑Tersenghi, ¡R. ¡ Zecchina, ¡JSTAT ¡111, ¡505 ¡(2003) ¡ ( n i 1 + n j 1 + n k 1 ) mod2 = J 1 n i ( p ) = 0 , 1 . . . ( n i M + n j M + n k M ) mod2 = J M J p = 0 , 1 UNSAT ¡ SAT ¡ α s

  19. XORSAT ¡ log(N) ¡loops ¡ ( n i 1 + n j 1 + n k 1 ) mod2 = J 1 . . . ( n i M + n j M + n k M ) mod2 = J M #clusters = Σ M. ¡Mezard, ¡F. ¡Ricci-­‑Tersenghi, ¡R. ¡ Zecchina, ¡JSTAT ¡111, ¡505 ¡(2003) ¡ ∆ E = O ( N ) UNSAT ¡ Non-­‑clustered ¡ Clustered ¡ Supercooled ¡ Liquid ¡ α d Glass ¡ α s liquid ¡

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend