the gaussian hardy-littlewood maximal function

25
The Gaussian Hardy-Littlewood Maximal Function YFAW ’14 Lancaster, UK Jonas Teuwen Delft Institute for Applied Mathematics Delft University of Technology, The Netherlands 24 April 2014

Upload: jonas-teuwen

Post on 20-Jul-2015

379 views

Category:

Science


0 download

TRANSCRIPT

Page 1: The Gaussian Hardy-Littlewood Maximal Function

The Gaussian Hardy-Littlewood MaximalFunction

YFAW ’14 Lancaster, UK

Jonas Teuwen

Delft Institute for Applied MathematicsDelft University of Technology, The Netherlands

24 April 2014

Page 2: The Gaussian Hardy-Littlewood Maximal Function

Introduction

Goal Build a satisfactory harmonic analytic theory for theGaussian measure

dγ(x) :=e−|x |

2

πd2

dx .

and the Ornstein-Uhlenbeck operator

L :=1

2∆− x · ∇

Just do it? Most theory relies on the doubling property of themeasure µ:

µ(B2r (x)) 6 Cµ(Br (x))

uniformly in r and x .

Page 3: The Gaussian Hardy-Littlewood Maximal Function

Introduction

Goal Build a satisfactory harmonic analytic theory for theGaussian measure

dγ(x) :=e−|x |

2

πd2

dx .

and the Ornstein-Uhlenbeck operator

L :=1

2∆− x · ∇

Just do it? Most theory relies on the doubling property of themeasure µ:

µ(B2r (x)) 6 Cµ(Br (x))

uniformly in r and x .

Page 4: The Gaussian Hardy-Littlewood Maximal Function

As you might guess. . .

. . . the Gaussian measure is non-doubling(but. . . maybe local?)

However, there is a kind of local doubling property! First this, wedefine the admissible balls

Ba := {B(x , r) : r 6 m(x)},

where,

m(x) := min

{1,

1

|x |

}.

For our admissible balls we then get the following lemma (Mauceri& Meda):

LemmaFor all Br (x) ∈ Ba we have that

γ(B2r (x)) 6 Cγ(Br (x)).

Page 5: The Gaussian Hardy-Littlewood Maximal Function

As you might guess. . .

. . . the Gaussian measure is non-doubling(but. . . maybe local?)

However, there is a kind of local doubling property! First this, wedefine the admissible balls

Ba := {B(x , r) : r 6 m(x)},

where,

m(x) := min

{1,

1

|x |

}.

For our admissible balls we then get the following lemma (Mauceri& Meda):

LemmaFor all Br (x) ∈ Ba we have that

γ(B2r (x)) 6 Cγ(Br (x)).

Page 6: The Gaussian Hardy-Littlewood Maximal Function

Some modifications – Gaussian conesGaussian harmonic analysis is local in the way that we use a cut-offcone for our non-tangential maximal function

Γ(A,a)x (γ) := {(x , y) ∈ R2d : |x − y | < At and t 6 am(x)}.

−1

1−1

1

1

x

y

t

Figure: This is a cut-off cone!

Page 7: The Gaussian Hardy-Littlewood Maximal Function

Gaussian cones – Useful consequences

1. On the cone Γ(A,a)x (γ) we have t|x | 6 aA,

2. If |x − y | < At and t 6 am(x) then |x | ∼ |y |.

Using these two we have

1. If we have this then additionally t|y | . 1

2. e−|x |2 ∼ e−|y |

2.

Furthermore we define the annuli (Ck)k>0 through

Ck(x) :=

{2Bt(x) if k = 0,

2k+1Bt(x) \ 2kBt(x) if k > 1.

Page 8: The Gaussian Hardy-Littlewood Maximal Function

Gaussian cones – Useful consequences

1. On the cone Γ(A,a)x (γ) we have t|x | 6 aA,

2. If |x − y | < At and t 6 am(x) then |x | ∼ |y |.

Using these two we have

1. If we have this then additionally t|y | . 1

2. e−|x |2 ∼ e−|y |

2.

Furthermore we define the annuli (Ck)k>0 through

Ck(x) :=

{2Bt(x) if k = 0,

2k+1Bt(x) \ 2kBt(x) if k > 1.

Page 9: The Gaussian Hardy-Littlewood Maximal Function

Gaussian cones – Useful consequences

1. On the cone Γ(A,a)x (γ) we have t|x | 6 aA,

2. If |x − y | < At and t 6 am(x) then |x | ∼ |y |.

Using these two we have

1. If we have this then additionally t|y | . 1

2. e−|x |2 ∼ e−|y |

2.

Furthermore we define the annuli (Ck)k>0 through

Ck(x) :=

{2Bt(x) if k = 0,

2k+1Bt(x) \ 2kBt(x) if k > 1.

Page 10: The Gaussian Hardy-Littlewood Maximal Function

Only one theorem. . . – The Euclidean case

TheoremLet u ∈ C∞c (Rd), then

sup(y ,t)∈Rd+1

+

|x−y |<t

|et2∆u(y)| . supr>0

1

|Br (x)|

∫Br (x)

|u|dλ︸ ︷︷ ︸Mu(x)

where λ is the Lebesgue measure.

The proof is straightforward, as et∆ is a convolution-type operator.So et∆ = ρt ∗ u where

ρt(ξ) :=e−|ξ|

2/4t

(4πt)d2

.

is the heat kernel.(There are many theorems about such convolution-type operators)

Page 11: The Gaussian Hardy-Littlewood Maximal Function

How to (ad hoc) prove it?

Let Ck := 2k+1B \ 2kB as before then

|et2∆u(y)| 6 1

(4πt2)d2

∫Rd

e−|y−ξ|2/4t2 |u(ξ)| dξ

61

(4πt2)d2

∞∑k=0

e−c4k∫Ck (Bt(x))

|u(ξ)| dξ

61

(4πt2)d2

∞∑k=0

e−c4k |B2k+1t(x)|Mu(y)

. Mu(y)1

td

∞∑k=0

e−c4k td2kd .

. Mu(y).

taking the supremum, and we are done.

Page 12: The Gaussian Hardy-Littlewood Maximal Function

Intermezzo: The Ornstein-Uhlenbeck semigroup

Remember the Ornstein-Uhlenbeck operator L?

L :=1

2∆− x · ∇.

L has the associated Ornstein-Uhlenbeck semigroup etL which inits turn has an associated Schwartz kernel:

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ)dξ, u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =1

πd2

1

(1− e−2t)d2

exp

(−|e

−tx − ξ|2

1− e−2t

).

Not nearly as convenient to work with as the heat kernel!

Page 13: The Gaussian Hardy-Littlewood Maximal Function

Intermezzo: The Ornstein-Uhlenbeck semigroup

Remember the Ornstein-Uhlenbeck operator L?

L :=1

2∆− x · ∇.

L has the associated Ornstein-Uhlenbeck semigroup etL which inits turn has an associated Schwartz kernel:

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ)dξ, u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =1

πd2

1

(1− e−2t)d2

exp

(−|e

−tx − ξ|2

1− e−2t

).

Not nearly as convenient to work with as the heat kernel!

Page 14: The Gaussian Hardy-Littlewood Maximal Function

Same ad hoc proof?

What goes wrong when we bluntly replace ∆ by L and theLebesgue measure by the Gaussian?

On Ck we have a lower bound for |x − ξ|, so,

|e−tx − ξ| > |x − ξ| − (1− e−t)|x |> |x − ξ| − t|x |.

Here the cone condition t|x | . 1 comes into play. Still, this is a bitunsatisfactory. We require a Gaussian measure. So what can bedone?

Page 15: The Gaussian Hardy-Littlewood Maximal Function

Same ad hoc proof?

What goes wrong when we bluntly replace ∆ by L and theLebesgue measure by the Gaussian?On Ck we have a lower bound for |x − ξ|, so,

|e−tx − ξ| > |x − ξ| − (1− e−t)|x |> |x − ξ| − t|x |.

Here the cone condition t|x | . 1 comes into play. Still, this is a bitunsatisfactory. We require a Gaussian measure. So what can bedone?

Page 16: The Gaussian Hardy-Littlewood Maximal Function

Unsatisfactory? – Some observations

1. We use the Lebesgue measure again, while we want aGaussian theory. Perspective,. . .

2. In the end we want all admissibility paramaters and aperturesa and A. Proof gets very messy (e.g., Urbina and Pineda),

3. Simple observations shows that the kernel should besymmetric in its arguments against the Gaussian measure.

So, honouring these observations we come to. . .

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ) γ(dξ), u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =

exp

(−e−2t |x − ξ|2

1− e−2t

)(1− e−t)

d2

exp

(−2e−t

〈x , ξ〉1 + e−t

)(1 + e−t)

d2

.

Page 17: The Gaussian Hardy-Littlewood Maximal Function

Unsatisfactory? – Some observations

1. We use the Lebesgue measure again, while we want aGaussian theory. Perspective,. . .

2. In the end we want all admissibility paramaters and aperturesa and A. Proof gets very messy (e.g., Urbina and Pineda),

3. Simple observations shows that the kernel should besymmetric in its arguments against the Gaussian measure.

So, honouring these observations we come to. . .

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ) γ(dξ), u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =

exp

(−e−2t |x − ξ|2

1− e−2t

)(1− e−t)

d2

exp

(−2e−t

〈x , ξ〉1 + e−t

)(1 + e−t)

d2

.

Page 18: The Gaussian Hardy-Littlewood Maximal Function

Unsatisfactory? – Some observations

1. We use the Lebesgue measure again, while we want aGaussian theory. Perspective,. . .

2. In the end we want all admissibility paramaters and aperturesa and A. Proof gets very messy (e.g., Urbina and Pineda),

3. Simple observations shows that the kernel should besymmetric in its arguments against the Gaussian measure.

So, honouring these observations we come to. . .

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ) γ(dξ), u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =

exp

(−e−2t |x − ξ|2

1− e−2t

)(1− e−t)

d2

exp

(−2e−t

〈x , ξ〉1 + e−t

)(1 + e−t)

d2

.

Page 19: The Gaussian Hardy-Littlewood Maximal Function

Unsatisfactory? – Some observations

1. We use the Lebesgue measure again, while we want aGaussian theory. Perspective,. . .

2. In the end we want all admissibility paramaters and aperturesa and A. Proof gets very messy (e.g., Urbina and Pineda),

3. Simple observations shows that the kernel should besymmetric in its arguments against the Gaussian measure.

So, honouring these observations we come to. . .

etLu(x) =

∫Rd

Mt(x , ξ)u(ξ) γ(dξ), u ∈ C∞c (Rd)

where the Mehler kernel Mt is given by

Mt(x , ξ) =

exp

(−e−2t |x − ξ|2

1− e−2t

)(1− e−t)

d2

exp

(−2e−t

〈x , ξ〉1 + e−t

)(1 + e−t)

d2

.

Page 20: The Gaussian Hardy-Littlewood Maximal Function

Estimating the Mehler kernel

On Ck this is now easier, for t . 1 and t|x | . 1:

Mt2(x , ξ) 6e−c4k

(1− e−2t2)d2

exp

(−2e−t

2 〈x , ξ〉1 + e−t2

)

6e−c4k

(1− e−2t2)d2

exp(−|〈x , ξ〉|)

6e−c4k

(1− e−2t2)d2

exp(|〈x , ξ − x〉|)e|x |2

6e−c4k

(1− e−2t2)d2

exp(2k+1t|x |)e|x |2

Page 21: The Gaussian Hardy-Littlewood Maximal Function

Was that enough? – Putting the things together

Let Ck := 2k+1B \ 2kB as before then

|et2Lu(y)| 6 1

(1− e−2t2)d2

∞∑k=0

e−c4k e2k+1t|y |e|y |2∫Ck (Bt(x))

|u| dγ

6 Mγu(y)e|y |

2

(1− e−2t2)d2

∞∑k=0

e−c4k e2k+1t|y |γ(B2k+1t(x)).

Estimating Gaussian balls:

γ(B2k+1t(x)) .d 2d(k+1)tde2k+2t|x |e−|x |2.

Page 22: The Gaussian Hardy-Littlewood Maximal Function

Was that enough? – Putting the things together

Let Ck := 2k+1B \ 2kB as before then

|et2Lu(y)| 6 1

(1− e−2t2)d2

∞∑k=0

e−c4k e2k+1t|y |e|y |2∫Ck (Bt(x))

|u| dγ

6 Mγu(y)e|y |

2

(1− e−2t2)d2

∞∑k=0

e−c4k e2k+1t|y |γ(B2k+1t(x)).

Estimating Gaussian balls:

γ(B2k+1t(x)) .d 2d(k+1)tde2k+2t|x |e−|x |2.

Page 23: The Gaussian Hardy-Littlewood Maximal Function

We do need to use the locality

As before, locally:

1. |x | ∼ |y |2. t|x | . 1 and t|y | . 1.

Combining we neatly get

|et2Lu(y)| . Mγu(y)td

(1− e−2t2)d2

∞∑k=0

e−c4k eC2k 2dk

Which is bounded for bounded t.(Thanks for your attention!)

Page 24: The Gaussian Hardy-Littlewood Maximal Function

We do need to use the locality

As before, locally:

1. |x | ∼ |y |2. t|x | . 1 and t|y | . 1.

Combining we neatly get

|et2Lu(y)| . Mγu(y)td

(1− e−2t2)d2

∞∑k=0

e−c4k eC2k 2dk

Which is bounded for bounded t.

(Thanks for your attention!)

Page 25: The Gaussian Hardy-Littlewood Maximal Function

We do need to use the locality

As before, locally:

1. |x | ∼ |y |2. t|x | . 1 and t|y | . 1.

Combining we neatly get

|et2Lu(y)| . Mγu(y)td

(1− e−2t2)d2

∞∑k=0

e−c4k eC2k 2dk

Which is bounded for bounded t.(Thanks for your attention!)