20.07.2013 Views

Estimation optimale du gradient du semi-groupe de la chaleur sur le ...

Estimation optimale du gradient du semi-groupe de la chaleur sur le ...

Estimation optimale du gradient du semi-groupe de la chaleur sur le ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

K. Koufany, G. Zhang / Journal of Functional Analysis 236 (2006) 546–580 567<br />

where H β is the component of H given by<br />

H β = <br />

[Eβ,vα]¯vα<br />

α∈Ψ + n<br />

and Eβ is the root vector of β.<br />

Now we fix for the rest of this section β ∈ Ψ +,(1)<br />

c such that<br />

β| t −<br />

2<br />

C = γj − γj−1<br />

The root vector Eβ has the form Eβ = D(cj , ¯w) with w = ej,j−1 or w = ej−1,j being one of the<br />

basis vectors {vα}. Observe that [Eβ,vα]=0 un<strong>le</strong>ss α is in the set Ψ1 ∪ Ψ2 ∪ Ψ3 where<br />

<br />

Ψ1 = α ∈ Ψ + n : α| t −<br />

<br />

Ψ2 = α ∈ Ψ + n : α| t −<br />

2<br />

C = γk + γj−1<br />

2<br />

C = γk + γj−1<br />

<br />

Ψ3 = α ∈ Ψ + n : α| t −<br />

<br />

γj−1<br />

= .<br />

C 2<br />

Consi<strong>de</strong>r the Poincaré–Birkhoff–Witt <strong>de</strong>composition<br />

and <strong>le</strong>t π be the projection<br />

.<br />

<br />

,k j − 1 ,<br />

<br />

,k j ,<br />

U(gC) = U(gC)kC + U(aC + n −<br />

C )<br />

π : U(gC) = U(gC)kC + U(aC + n −<br />

C ) → U(aC + n −<br />

C ).<br />

The function f is now viewed as a function on G = NAK, and the group A will be i<strong>de</strong>ntified<br />

as (R + ) r . Un<strong>de</strong>r this i<strong>de</strong>ntification, f satisfies, furthermore, the equation<br />

R π H β f = 0,<br />

where R is the mapping from U(aC + nC) to differential operators on NA <strong>de</strong>fined by<br />

∂<br />

R(ξk) = tk , R(X−α) = t<br />

∂tk<br />

α X−α,<br />

for ξk ∈ a, 1 k r, and X−α ∈ n i<strong>de</strong>ntified with the corresponding <strong>le</strong>ft-invariant differential<br />

operator.<br />

We will prove that operator t − 1 2 (βj −βj−1) R(π(H β )) has analytic coefficient near t = 0 and<br />

study the in<strong>du</strong>ced equation of t − 1 2 (βj −βj−1) R(π(H β ))f = 0.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!