a course in interacting particle systemsstaff.utia.cas.cz/swart/lecture_notes/introsemin.pdf ·...
TRANSCRIPT
![Page 1: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/1.jpg)
A Course in Interacting Particle Systems
Jan M. Swart
Lecture 1: Introduction
Jan M. Swart Particle Systems
![Page 2: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/2.jpg)
Interacting Particle Systems
I Interacting particle systems are mathematical models forcollective behavior.
I Applications in physics (atoms & molecules), biology(organisms) & sociology, financial mathematics (people).
I Simple rules lead to complicated behavior.
I Markovian dynamics.
I Easy to simulate, but not always easy to prove; open problems.
I Rigorous methods lead to better understanding.
Jan M. Swart Particle Systems
![Page 3: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/3.jpg)
Interacting Particle Systems
Interacting Particle Systems are continuous-time Markov processesX = (Xt)t≥0 with state space of the form SΛ, where:
I S is a finite set, called the local state space.
I Λ is a countable set, called the lattice.
We denote an element x ∈ SΛ as
x =(x(i)
)i∈Λ
with x(i) ∈ S ∀ i ∈ Λ.
We call Xt(i) the local state of the process at time t and positioni ∈ Λ.
Jan M. Swart Particle Systems
![Page 4: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/4.jpg)
Generator and local maps
The generator G of an interacting particle system can be written inthe form
Gf (x) =∑m∈G
rm{
f(m(x)
)− f(x)},
where (rm)m∈G are nonnegative rates and G is a collection of localmaps m : SΛ → SΛ.Poisson construction: The process can be constructed byapplying each map m ∈ G at the times of a Poisson process withintensity rm.Generator construction: If Λ is finite, then the transitionprobabilities Pt(x , y) are given by
Pt := e tG :=∞∑n=0
1
n!(tG )n.
Jan M. Swart Particle Systems
![Page 5: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/5.jpg)
Examples of lattices
Often, the lattice is a graph (Λ,E ) with (undirected) edge set E .We denote the corresponding set of directed edges by:
E :={
(i , j) : {i , j} ∈ E}
Z2 T2
Jan M. Swart Particle Systems
![Page 6: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/6.jpg)
Examples of lattices
In particular, we equip Zd with the following edge sets:
Ed :={{i , j} : ‖i − j‖1 = 1
},
EdR :=
{{i , j} : 0 < ‖i − j‖∞ ≤ R
}.
We let Ed and EdR denote the corresponding directed edges and let
Ni resp. NRi := {j ∈ Zd : {i , j} ∈ Ed resp. Ed
R}
denote the neighborhood of i .
Jan M. Swart Particle Systems
![Page 7: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/7.jpg)
The voter model
For each i , j ∈ Λ, the voter model map votij : SΛ → SΛ is definedas
votij(x)(k) :=
{x(i) if k = j ,
x(k) otherwise.
In words, this copies the state of i onto j . The nearest neighborvoter model is defined by
Gvotf (x) =1
|N0|∑
(i ,j)∈Ed
{f(votij(x)
)− f(x)}
(x ∈ SΛ).
Similarly, NRi gives the range R voter model.
Jan M. Swart Particle Systems
![Page 8: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/8.jpg)
The voter model
With rate one, the site j adopts the type x(i) of a randomlychosen neighbor.
Interpretation 1 Sites are people, types are political parties; atrate one, people ask their neighbor whom to vote for.
Interpretation 2 Sites are organisms, types are genetic types; atrate one, an organism dies and is replaced by a clone of a randomlychosen neighbor.
Jan M. Swart Particle Systems
![Page 9: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/9.jpg)
The voter model
Time t = 0.
Jan M. Swart Particle Systems
![Page 10: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/10.jpg)
The voter model
Time t = 0.25.
Jan M. Swart Particle Systems
![Page 11: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/11.jpg)
The voter model
Time t = 0.5.
Jan M. Swart Particle Systems
![Page 12: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/12.jpg)
The voter model
Time t = 1.
Jan M. Swart Particle Systems
![Page 13: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/13.jpg)
The voter model
Time t = 2.
Jan M. Swart Particle Systems
![Page 14: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/14.jpg)
The voter model
Time t = 4.
Jan M. Swart Particle Systems
![Page 15: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/15.jpg)
The voter model
Time t = 8.
Jan M. Swart Particle Systems
![Page 16: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/16.jpg)
The voter model
Time t = 16.
Jan M. Swart Particle Systems
![Page 17: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/17.jpg)
The voter model
Time t = 31.25.
Jan M. Swart Particle Systems
![Page 18: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/18.jpg)
The voter model
Time t = 62.5.
Jan M. Swart Particle Systems
![Page 19: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/19.jpg)
The voter model
Time t = 125.
Jan M. Swart Particle Systems
![Page 20: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/20.jpg)
The voter model
Time t = 250.
Jan M. Swart Particle Systems
![Page 21: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/21.jpg)
The voter model
Time t = 500.
Jan M. Swart Particle Systems
![Page 22: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/22.jpg)
The voter model
The behavior of the voter model strongly depends on thedimension.
Clustering in dimensions d = 1, 2.
Stable behavior in dimensions d ≥ 3.
Jan M. Swart Particle Systems
![Page 23: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/23.jpg)
The voter model
Cut of 3-dimensional model, time t = 0.
Jan M. Swart Particle Systems
![Page 24: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/24.jpg)
The voter model
Cut of 3-dimensional model, time t = 1.
Jan M. Swart Particle Systems
![Page 25: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/25.jpg)
The voter model
Cut of 3-dimensional model, time t = 2.
Jan M. Swart Particle Systems
![Page 26: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/26.jpg)
The voter model
Cut of 3-dimensional model, time t = 4.
Jan M. Swart Particle Systems
![Page 27: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/27.jpg)
The voter model
Cut of 3-dimensional model, time t = 8.
Jan M. Swart Particle Systems
![Page 28: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/28.jpg)
The voter model
Cut of 3-dimensional model, time t = 16.
Jan M. Swart Particle Systems
![Page 29: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/29.jpg)
The voter model
Cut of 3-dimensional model, time t = 32.
Jan M. Swart Particle Systems
![Page 30: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/30.jpg)
The voter model
Cut of 3-dimensional model, time t = 64.
Jan M. Swart Particle Systems
![Page 31: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/31.jpg)
The voter model
Cut of 3-dimensional model, time t = 125.
Jan M. Swart Particle Systems
![Page 32: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/32.jpg)
The voter model
Cut of 3-dimensional model, time t = 250.
Jan M. Swart Particle Systems
![Page 33: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/33.jpg)
The Contact Process
Let S = {0, 1} with 0 = empty and 1 = occupied.
For each i , j ∈ Λ, we define a branching mapbraij : {0, 1}Λ → {0, 1}Λ as
braij(x)(k) :=
{x(i) ∨ x(j) if k = j ,
x(k) otherwise.
For each i ∈ Λ, we also define a death mapdeathi : {0, 1}Λ → {0, 1}Λ as
deathi (x)(k) :=
{0 if k = i ,
x(k) otherwise.
Jan M. Swart Particle Systems
![Page 34: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/34.jpg)
The Contact Process
The nearest neighbor contact process with infection rate λ isdefined by the generator
Gcontf (x) :=λ∑
(i ,j)∈Ed
{f ((braij(x))− f
(x)}
+∑i∈Zd
{f ((deathi (x))− f
(x)}
(x ∈ {0, 1}Zd).
Interpretation 1 1 = infected, 0 = healthy, sites infect eachneighbor with rate λ and recover with rate one.
Interpretation 2 1 = occupied, 0 = empty, sites place offspring oneach neighboring site with rate λ and die with rate one.
Jan M. Swart Particle Systems
![Page 35: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/35.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 0.
Jan M. Swart Particle Systems
![Page 36: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/36.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 1.
Jan M. Swart Particle Systems
![Page 37: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/37.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 2.
Jan M. Swart Particle Systems
![Page 38: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/38.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 3.
Jan M. Swart Particle Systems
![Page 39: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/39.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 4.
Jan M. Swart Particle Systems
![Page 40: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/40.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 5.
Jan M. Swart Particle Systems
![Page 41: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/41.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 6.
Jan M. Swart Particle Systems
![Page 42: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/42.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 7.
Jan M. Swart Particle Systems
![Page 43: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/43.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 8.
Jan M. Swart Particle Systems
![Page 44: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/44.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 9.
Jan M. Swart Particle Systems
![Page 45: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/45.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 10.
Jan M. Swart Particle Systems
![Page 46: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/46.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 11.
Jan M. Swart Particle Systems
![Page 47: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/47.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 12.
Jan M. Swart Particle Systems
![Page 48: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/48.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 13.
Jan M. Swart Particle Systems
![Page 49: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/49.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 14.
Jan M. Swart Particle Systems
![Page 50: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/50.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 15.
Jan M. Swart Particle Systems
![Page 51: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/51.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 16.
Jan M. Swart Particle Systems
![Page 52: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/52.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 17.
Jan M. Swart Particle Systems
![Page 53: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/53.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 18.
Jan M. Swart Particle Systems
![Page 54: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/54.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 19.
Jan M. Swart Particle Systems
![Page 55: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/55.jpg)
The contact process
Contact process with infection rate λ = 2 and death rate d = 1.Time t = 20.
Jan M. Swart Particle Systems
![Page 56: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/56.jpg)
The Contact Process
λ
θ(λ)
0 1 2 3 4
0.25
0.5
0.75
1
λc
The survival probability
θ(λ) := P1{0}[Xt 6= 0 ∀t ≥ 0
]exhibits a phase transition. In one dimenson λc ≈ 1.649.
Jan M. Swart Particle Systems
![Page 57: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/57.jpg)
Ising and Potts models
Let S be any finite set. For any x ∈ SΛ and i ∈ Λ, let
Nx ,i (σ) :=∑j∈Ni
1{x(j) = σ} (σ ∈ S)
denote the number of neighbors of site i that have the spin valueσ ∈ S . In the Potts model with Glauber dynamics,
site i flips to the value σ with rateeβNx,i (σ)∑τ∈S eβNx,i (τ)
.
I.e., update with rate one, choose new value σ proportional toeβNx,i (σ). For β ≥ 0 ferromagnetic.
Jan M. Swart Particle Systems
![Page 58: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/58.jpg)
The Potts model
β = 1.2, time t = 0.
Jan M. Swart Particle Systems
![Page 59: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/59.jpg)
The Potts model
β = 1.2, time t = 1.
Jan M. Swart Particle Systems
![Page 60: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/60.jpg)
The Potts model
β = 1.2, time t = 2.
Jan M. Swart Particle Systems
![Page 61: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/61.jpg)
The Potts model
β = 1.2, time t = 4.
Jan M. Swart Particle Systems
![Page 62: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/62.jpg)
The Potts model
β = 1.2, time t = 8.
Jan M. Swart Particle Systems
![Page 63: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/63.jpg)
The Potts model
β = 1.2, time t = 16.
Jan M. Swart Particle Systems
![Page 64: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/64.jpg)
The Potts model
β = 1.2, time t = 32.
Jan M. Swart Particle Systems
![Page 65: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/65.jpg)
The Potts model
β = 1.2, time t = 64.
Jan M. Swart Particle Systems
![Page 66: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/66.jpg)
The Potts model
β = 1.2, time t = 125.
Jan M. Swart Particle Systems
![Page 67: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/67.jpg)
The Potts model
β = 1.2, time t = 250.
Jan M. Swart Particle Systems
![Page 68: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/68.jpg)
The Potts model
β = 1.2, time t = 500.
Jan M. Swart Particle Systems
![Page 69: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/69.jpg)
The Ising model
This looks superficially like the voter model, but:
I Even in large clusters, single sites can still flip to other colors.
I Clustering happens only for β above a critical value βc.
I 0 < βc <∞ in dimensions d ≥ 2 but βc =∞ in dimensionone.
Jan M. Swart Particle Systems
![Page 70: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/70.jpg)
A one-dimensional Potts model
space
time
0 100 200 300 400 500
0
500
1000
1500
2000
2500
In one-dimensional Potts models, the cluster size remainsbounded in time even at very high β (= low temperature).
Jan M. Swart Particle Systems
![Page 71: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/71.jpg)
The voter model
space
time
0 100 200 300 400 500
0
500
1000
1500
2000
2500
A one-dimensional voter model.
Jan M. Swart Particle Systems
![Page 72: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/72.jpg)
The Ising model
The Ising model is the Potts model with S = {−1,+1}.
On a large, but finite lattice Λ, freeze the boundary spins to +1.
Since Λ is finite, there is a unique invariant law ν+.The spontaneous magnetization is the function
m∗(β) := limΛ↑Zd
∫ν+(dx) x(0).
One has clustering, i.e., long-range order, iff m∗(β) > 0.
Jan M. Swart Particle Systems
![Page 73: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/73.jpg)
The magnetization of the Ising model
β
m∗(β)
0 0.5 1 1.5
0.25
0.5
0.75
1
βc
Onsager (1944) proved that for the model on Z2,
m∗(β) =
{ (1− sinh(β)−4
)1/8for β ≥ βc := log(1 +
√2),
0 for β ≤ βc.
Jan M. Swart Particle Systems
![Page 74: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/74.jpg)
The magnetization of the Ising model
β
m∗(β)
1
βc
For the model on Z3, it is known that m∗ is continuous,nondecreasing in β, and there exists a 0 < βc <∞ such thatm∗(β) = 0 for β ≤ βc while m∗(β) > 0 for β > βc. Continuity atβc proved by Aizenman, Duminil-Copin & Sidoravicius (2014).
Jan M. Swart Particle Systems
![Page 75: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/75.jpg)
Phase transitions
The contact process goes through a phase transition at λc.The Ising model goes through a phase transition at βc.
In both cases, below the critical point (λc resp. βc), the system isin a phase where there is a unique invariant law and absence oflong-range order.
Above λc resp. βc, there are multiple invariant laws and long-rangeorder.
Jan M. Swart Particle Systems
![Page 76: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/76.jpg)
Phase transitions
Since the order parameter θ(λ) resp. m∗(β) is continuous at thecritical point, the phase transitions of the contact process and Isingmodel are second order or continuous phase transitions.
By contrast, for Potts models with a high number of colors (indimension two, for q := |S | > 4), the analogue of themagnetization m∗ has a jump at βc. These systems have q orderedinvariant laws for β > βc, one unordered invariant law for β < βc,while at β = βc, all q + 1 invariant laws coexist (q ordered statesand one disordered state). Such a phase transition is called firstorder.
Jan M. Swart Particle Systems
![Page 77: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/77.jpg)
Critical exponents
β
m∗(β)
1
βc
Nonrigorous renormalization group theory explains that
m∗(β) ∝ (β − βc)c as β ↓ βc,
where the critical exponent c is given by
c = 1/8 in dim 2, c ≈ 0.326 in dim 3, and c = 1/2 in dim ≥ 4.
Jan M. Swart Particle Systems
![Page 78: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/78.jpg)
Critical exponents
For the contact process, it is believed that
θ(λ) ∝ (λ− λc)c as β ↓ βc,
where the critical exponent c is given by
c ≈ 0.276 in dim 1, c ≈ 0.583 in dim 2,
c ≈ 0.813 in dim 3, and c = 1 in dim ≥ 4.
There are other critical exponents associated with other quantities(such as the correlation length) or with the power-law decay ofcorrelations at the critical point.
Jan M. Swart Particle Systems
![Page 79: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/79.jpg)
Critical exponents
Critical exponents are believed to be universal. For example, forthe range R model, the value of the critical point depends on Rbut the critical exponent does not. Critical exponents associatedwith the d = 3 Ising model have even been measured in realphysical systems.
Critical exponents and more generally critical behavior areassociated only with second order phase transitions, and for thisreason physicists use the word “critical point” only for secondorder phase transitions.
Jan M. Swart Particle Systems
![Page 80: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/80.jpg)
Critical exponents
Critical exponents are explained by nonrigorous renormalizationgroup theory but so far, there is no general mathematical theory.Powerlaw behavior with well-defined critical exponents has beenproved in some special cases:
I In sufficiently high dimensions by means of the lace expansion.
I In a few exactly solvable models like the Ising model on Z2.
I In certain 2-dimensional models related to conformal fieldtheory and the Schramm Loewner Equation.
Jan M. Swart Particle Systems
![Page 81: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/81.jpg)
Variations on the voter model
The biased voter model with bias s ≥ 0 is the interacting particlesystem with state space {0, 1}Zd
and generator
Gbiasf (x) :=1
|Ni |∑
(i ,j)∈Ed
{f(votij(x))− f
(x)}
+s
|Ni |∑
(i ,j)∈Ed
{f(braij(x))− f
(x)}.
The paremeter s > 0 gives type 1 a (small) advantage.
Contrary to the voter model, even if we start with just a singleperson of type 1, there is a positive probability that type 1 neverdies out.
Models spread of new idea or technology, or advantageousmutation in biology.
Jan M. Swart Particle Systems
![Page 82: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/82.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 0 .
Jan M. Swart Particle Systems
![Page 83: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/83.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 10.
Jan M. Swart Particle Systems
![Page 84: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/84.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 20.
Jan M. Swart Particle Systems
![Page 85: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/85.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 30.
Jan M. Swart Particle Systems
![Page 86: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/86.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 40.
Jan M. Swart Particle Systems
![Page 87: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/87.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 50.
Jan M. Swart Particle Systems
![Page 88: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/88.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 60.
Jan M. Swart Particle Systems
![Page 89: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/89.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 70.
Jan M. Swart Particle Systems
![Page 90: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/90.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 80.
Jan M. Swart Particle Systems
![Page 91: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/91.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 90.
Jan M. Swart Particle Systems
![Page 92: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/92.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 100.
Jan M. Swart Particle Systems
![Page 93: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/93.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 110.
Jan M. Swart Particle Systems
![Page 94: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/94.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 120.
Jan M. Swart Particle Systems
![Page 95: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/95.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 130.
Jan M. Swart Particle Systems
![Page 96: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/96.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 140.
Jan M. Swart Particle Systems
![Page 97: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/97.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 150.
Jan M. Swart Particle Systems
![Page 98: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/98.jpg)
The biased voter model
Biased voter model with s = 0.2. Time t = 160.
Jan M. Swart Particle Systems
![Page 99: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/99.jpg)
The biased voter model
space
time
0 100 200 300 400 500
0
100
200
300
400
A one-dimensional biased voter model with bias s = 0.2.
Jan M. Swart Particle Systems
![Page 100: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/100.jpg)
The biased voter model
We can extend the biased voter model by also allowing deaths, i.e.,spontaneous jumps from 1 to 0.
This models the fact that complicated new ideas may be forgottenor organisms may die.
Whether 1’s have a positive probability to survive now depends ina nontrivial way on s and d .
In fact, the model appears to be in the same universality class asthe contact process.
Jan M. Swart Particle Systems
![Page 101: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/101.jpg)
The biased voter model
space
time
0 100 200 300 400 500
0
200
400
600
Process with bias s = 0.5, death rate d = 0.02.
Jan M. Swart Particle Systems
![Page 102: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/102.jpg)
A rebellious voter model
Consider a model with two types {0, 1} and let
fτ :=1
|Ni |∑j∈Ni
1{x(j) = τ}
be the frequency of type τ in the neighborhood Ni .
A person of type τ chooses a new type with rate
fτ + αf1−τ .
For α < 1, persons change their mind less often if they disagreewith a lot of neighbors.
As in a normal voter model, the probability that the newly chosentype is τ ′ is fτ ′ .
Used by Neuhauser & Pacala (1999) to model balancing selection.
Jan M. Swart Particle Systems
![Page 103: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/103.jpg)
A rebellious voter model
space
time
0 100 200 300 400 500
0
200
400
600
Process with α = 0.8 behaves more or less as a voter model.
Jan M. Swart Particle Systems
![Page 104: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/104.jpg)
A rebellious voter model
space
time
0 100 200 300 400 500
0
200
400
600
In the process with α = 0.3, cluster size remains bounded in time.
Jan M. Swart Particle Systems
![Page 105: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/105.jpg)
Coalescing random walks
Let Xt(i) = 1 (resp. 0) signify the presence (resp. absence) of a
particle and consider the maps rwij : {0, 1}Zd → {0, 1}Zd
rwijx(k) :=
0 if k = i ,
x(i) ∨ x(j) if k = j ,x(k) otherwise.
The process with generator
Grwf (x) :=1
|N0|∑
(i ,j)∈Ed
{f(rwijx
)− f(x)}
describes coalescing random walks.
Jan M. Swart Particle Systems
![Page 106: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/106.jpg)
Coalescing random walks
space
time
0 50 100 150 200
0
100
200
300
400
Jan M. Swart Particle Systems
![Page 107: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/107.jpg)
Reaction diffusion models
We can also add other maps to the dynamics, like thebranching map
braijx(k) :=
{x(i) ∨ x(j) if k = j ,
x(k) otherwise,
or even cooperative branching
coopii ′jx(k) :=
{ (x(i) ∧ x(i ′)
)∨ x(j) if k = j ,
x(k) otherwise.
Jan M. Swart Particle Systems
![Page 108: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/108.jpg)
Branching and coalescing random walks
space
time
0 50 100 150 200
0
100
200
300
400
Jan M. Swart Particle Systems
![Page 109: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/109.jpg)
Cooperative branching and coalescence
space
time
0 50 100 150 200
0
100
200
300
400
Cooperative branching rate 2.2.
Jan M. Swart Particle Systems
![Page 110: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/110.jpg)
Cooperative branching
space
time
0 50 100 150 200
0
100
200
300
400
Cooperative branching rate 3.
Jan M. Swart Particle Systems
![Page 111: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/111.jpg)
A cancellative system
Two more maps of interest are the annihilating random walk map
annijx(k) :=
0 if k = i ,
x(i) + x(j) mod(2) if k = j ,x(k) otherwise,
and the annihilating branching map
branijx(k) :=
{x(i) + x(j) mod(2) if k = j ,
x(k) otherwise,
Jan M. Swart Particle Systems
![Page 112: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/112.jpg)
A cancellative system
space
time
0 50 100 150 200
0
100
200
300
400
Annihilating random walks.
Jan M. Swart Particle Systems
![Page 113: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/113.jpg)
A cancellative system
space
time
0 50 100 150 200
0
10
20
30
40
A system of branching annihilating random walks.
Jan M. Swart Particle Systems
![Page 114: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/114.jpg)
Killing
Define a killing map as
killijx(k) :=
{ (1− x(i)
)∧ x(j) if k = j ,
x(k) otherwise,
which says that the particle at i , if present, kills any particle at j .
Jan M. Swart Particle Systems
![Page 115: A Course in Interacting Particle Systemsstaff.utia.cas.cz/swart/lecture_notes/IntroSemin.pdf · 2015. 10. 8. · Interacting Particle Systems I Interacting particle systems are mathematical](https://reader034.vdocuments.net/reader034/viewer/2022051604/5ff8165d69d40830c541a576/html5/thumbnails/115.jpg)
Branching and killing
space
time
0 50 100 150 200
0
100
200
300
400
A system with branching and killing.
Jan M. Swart Particle Systems