Multi-dimensional Brownian Motion with Darning Shuwen Lou - - PowerPoint PPT Presentation

multi dimensional brownian motion with darning
SMART_READER_LITE
LIVE PREVIEW

Multi-dimensional Brownian Motion with Darning Shuwen Lou - - PowerPoint PPT Presentation

Examples of darning processes Small time heat kernel estimate Multi-dimensional Brownian Motion with Darning Shuwen Lou University of Washington June 18, 2012 Shuwen Lou Multi-dimensional Brownian Motion with Darning Examples of darning


slide-1
SLIDE 1

Examples of darning processes Small time heat kernel estimate

Multi-dimensional Brownian Motion with Darning

Shuwen Lou

University of Washington

June 18, 2012

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-2
SLIDE 2

Examples of darning processes Small time heat kernel estimate

Outline

  • Motivation: What is a “darning" process?
  • Examples of darning processes
  • Main results: heat kernel estimates of multi-dimensional Brownian

motion with darnin

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-3
SLIDE 3

Examples of darning processes Small time heat kernel estimate

Question What is a “darning" process? Answer We either

  • “patch" the boundaries of multiple processes together, or
  • “collapse" some part of the state space of a process to a singleton.

Examples and pictures of darning processes will be given soon.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-4
SLIDE 4

Examples of darning processes Small time heat kernel estimate

Question How to construct a darning process? Answer Very roughly speaking, in terms of Dirichlet forms.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-5
SLIDE 5

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Examples of darning processes

Example (Circular Brownion motion) Idea:“Gluing" the two endpoints of an absorbing Brownian motion on an interval I := (0, 1). The Dirichlet form of circular Brownian motion is        F =

  • u : u ∈ BL(I), u(0+) = u(1−)
  • ∩ L2(I),

E(u, v) = 1 2

  • I

u′(x)v′(x)dx, where BL(I) =

  • u : u is absolutely continuous on I with
  • I

(u′)2dx < ∞

  • .

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-6
SLIDE 6

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Picture of circular Brownian motion

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-7
SLIDE 7

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Examples of darning processes

Example (Brownian motion with a “knot") Idea: Identifying two points on R as a singleton.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-8
SLIDE 8

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

(Continued) Suppose we identify two points a and b on R. The Dirichlet form of such a process is    F =

  • u : u ∈ W1,2(R), u(a) = u(b)
  • ,

E(u, v) = 1 2

  • R

u′(x)v′(x)dx. Remark In the same way we may identify up to countably many non-accumulating points on R so that the “knot" has multiple loops.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-9
SLIDE 9

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Question Why do we have to define a multi-dimensional Brownian motion as a “darning process"? Answer Even for the simplest case of R2 ⊔ R, a standard 2-dimensional Brownian motion never hits a singleton!

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-10
SLIDE 10

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Solution By “collapsing" the closure of an open set to a single point, one gets a “2-dimensional Brownian Motion" that does hit a singleton, which is called a Brownian motion with darning.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-11
SLIDE 11

Examples of darning processes Small time heat kernel estimate Circular Brownian motion Brownian motion with a “knot"

Another picture of this process

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-12
SLIDE 12

Examples of darning processes Small time heat kernel estimate

Question What is heat kernel estimate? Answer A Markov process has a transition semigroup: Px(Xt ∈ ·) := Pt(x, ·), which is a probability measure for every fixed pair of (t, x). In many cases, it is very hard to give the explicit expression of Pt(x, ·). We usually denote the density of Pt(x, ·) by p(t, x, y). Our goal is to find a function f(t, x, y) such that there exists some constant C > 0 such that 1 C · f(t, x, y) ≍ p(t, x, y) ≍ C · f(t, x, y), for all t, x, y.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-13
SLIDE 13

Examples of darning processes Small time heat kernel estimate

Proposition (Small time heat kernel estimate) There exist constants C1, C2 > 0, such that p(t, x, y) ≍       

1 √te− C1|x−y|2

t

, |y|g ≤ 1;

1 t e− C2|x−y|2

t

, |y|g > 1, for all x ∈ R, y ∈ R2\Bǫ, t ∈ [0, 1].

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-14
SLIDE 14

Examples of darning processes Small time heat kernel estimate

Theorem (Small time heat kernel estimate) There exist constants Ci > 0, 3 ≤ i ≤ 5, such that for all t ∈ [0, 1], x, y ∈ R2\Bǫ, p(t, x, y) ≍             

1 √te−

C3|x−y|2 g t

+ 1

t

  • 1 ∧ |x|g

√t

1 ∧ |y|g

√t

  • e− C4|x−y|2

e t

, |x|g < 1, |y|g < 1;

1 t e−

C5|x−y|2 g t

,

  • therwise.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-15
SLIDE 15

Examples of darning processes Small time heat kernel estimate

Theorem (Large time heat kernel estimate) There exist constants C6 > 0 such that p(X)(t, x, y) ≍ 1 √te− C6|x−y|2

t

, t ∈ [0, ∞), x, y ∈ R.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-16
SLIDE 16

Examples of darning processes Small time heat kernel estimate

Theorem (Large time heat kernel estimate) There exists constant C7 > 0 such that p(X)(t, x, y) ≍ 1 √te−

C7|x−y|2 g t

, t > 1, x ∈ R, y ∈ R2.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-17
SLIDE 17

Examples of darning processes Small time heat kernel estimate

Theorem (Large time heat kernel estimate) There exist constants Ci > 0, 8 ≤ i ≤ 10, such that for all t > 1, x, y ∈ R2\Bǫ, p(X)(t, x, y) ≍         

1 t e−

C8|x−y|2 g t

+ 1

√te−

C9|x|2 g+|y|2 g t

, |x|g > √t, |y|g > √t;

1 √te−

C10|x−y|2 g t

,

  • therwise.

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-18
SLIDE 18

Examples of darning processes Small time heat kernel estimate

Other obtained results related to small time HKE

  • Hölder continuity of parabolic functions
  • Counter example to parabolic Harnack inequality
  • Case of multiple straight lines
  • Case of a “handle" attached to a plane

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-19
SLIDE 19

Examples of darning processes Small time heat kernel estimate

Picture of the case of multiple straight lines

Shuwen Lou Multi-dimensional Brownian Motion with Darning

slide-20
SLIDE 20

Examples of darning processes Small time heat kernel estimate

Picture of a “handle" attached to a plane

Shuwen Lou Multi-dimensional Brownian Motion with Darning