Non linear elliptic systems and families of quasiregular...
Transcript of Non linear elliptic systems and families of quasiregular...
Non linear elliptic systems and families ofquasiregular mappings
K.Astala, A.Clop, D.Faraco, J.Jääskelainen, L.Székelyhidi Jr
Santander , July 2015
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
The main character
The nonlinear Beltrami Equation
∂z f (z) = H(z , ∂z f (z))
Nonlinearity H : C× C→ C s.t(H1) H is k-Lipschitz in the second variable, k < 1
|H(z ,w1)−H(z ,w2)| ≤ |w1 − w2|,
for almost every z ∈ C, and the normalization H(z , 0) ≡ 0holds.
(H2) For every w ∈ C, z 7→ H(z ,w) is measurable on C,
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Examples
C linear Beltrami equation.
∂f (z) = µ∂f (z) ‖µ‖L∞ ≤ k < 1
R linear Beltrami
∂f (z) = µ∂f (z) + ν∂f (z), ‖µ‖∞ + ‖ν‖∞ ≤ k < 1
Governs linear elliptic systems.Autonomous equations
∂f (z) = H(∂f )
γ-equations. Let γ be a curve in the complex plane.
∂f (z) = µdist(γ, ∂f (z)) ‖µ‖∞ ≤ k < 1
Key to the solution of Tartar conjecture by F-Szekelyhidi Jr.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
History of nonlinear Beltrami equation
Introduced by Bojarksi and Iwaniec 70Astala, Iwaniec and Saksman (Beltrami Operators in the plane)Astala Iwaniec and Martin [Monograph]. Governs all planarelliptic systems. Good existence theory for H(z , f ,w)
Good example of using topology of the solutions to go beyondthe classical ellptic P.D.E techniques. e.g Solution to CalderónProblem
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Quasiconformal maps
Normalized solutions = Homeomorphism fixing 0 and 1
‖∂f (z)‖ ≤ k |∂f |︸ ︷︷ ︸distortion inequality
Recall: Homeomorphic functions satisfying the distortion inequalityare called quasiconformal.Not homeomorphic solutions of the distortions inequality are calledquasiregular mappings or mappings of bounded distortion.Quasiregular mappings factorize as f = h F where F is anormalized quasiconformal solution and h is holomorphic.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Uniqueness of Normalized solutions
Unique conformal mapping fixing 0 and 1 is z . what happens forBeltrami equations?
C linear Beltrami equation.
∂f (z) = µ∂f (z) |µ‖L∞ ≤ k < 1
Yes from StoilowGeneral R linear Beltrami
∂f (z) = µ∂f (z) + ν∂f (z), ‖µ‖∞ + ‖ν‖∞ ≤ k < 1
Yes Astala-Iwaniec-Martin 2006. Reduction to the reduced equation
∂f (z) = µRe(∂f (z)) |µ‖L∞ ≤ k < 1
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Uniqueness for normalized γ quasiconformal maps
Theorem Suppose 0 ∈ γ. If there are zi ∈ C, i = 1, 2, y q0 ∈ γsuch that ϕ(zi ) = q0zi , then ϕ(z) ≡ q0z .Corollary 1 If f ∈W 1,2
loc (C) is a global homeomorphism s.t
|∂f (z)| ≤ kdist(∂f (z), [0, 1])
and f (0) = 0, f (z) = z , for some z ⊂ C, f (z) ≡ z .Corollary 2In particular if 0, Id ∈ Γ and H : C× C→ C k Lipschitz is sucha− = H(z , a+) for almost every z and every A = (a+, a−) ∈ Γ thereis a unique normalized solution to
∂f (z) = H(z , ∂f (z))
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Uniqueness of normalized solutions for non linear systems
Is there always uniqueness of normalized solutions to nonlinearBeltrami equations?No! ACFJS IMRN 2011
Uniqueness if lim sup|z|→∞ K (z) = k(z)−1k(z)+1 <
√2
If g(z) = z Uniqueness holds if lim sup|z|→∞ K (z) ≤ 2There is uniqueness for autonomous equations.Different bound for H(f , ∂f ).Uniqueness if there is γ(t) : γ(0) = 0, γ(1) = 1 and such thatH(z , γ(t)) ∈ Lp0 p0 < 1.The first and second point are sharp up to the end point!
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Uniqueness of normalized solutions for non linear systems
Is there always uniqueness of normalized solutions to nonlinearBeltrami equations?No! ACFJS IMRN 2011
Uniqueness if lim sup|z|→∞ K (z) = k(z)−1k(z)+1 <
√2
If g(z) = z Uniqueness holds if lim sup|z|→∞ K (z) ≤ 2There is uniqueness for autonomous equations.Different bound for H(f , ∂f ).Uniqueness if there is γ(t) : γ(0) = 0, γ(1) = 1 and such thatH(z , γ(t)) ∈ Lp0 p0 < 1.The first and second point are sharp up to the end point!
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Difference of two H solutions is quasiregular
Let, f,g solve the same! H equation. Then
∂(f (z)− g(z)) = H(∂(f ))−H(∂(g))
Hence,
|∂(f (z)− g(z))| = |H(∂(f ))−H(∂(g))| ≤ k |∂(f (z)− g(z))|
i.e f − g is quasiregular.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Let 1 < K = k−1k+1 <∞. Then
deg(f − g) ≤ K 2
Proof: Follows from writing f − g = P(H) P an holomorphicpolynomial and H a normalized quasiconformal mapping
|z |dK ≤ |f − g | ≤ C |z |K
Corollary if K 2 < 2, d = 1 and thus f = g .In the case, g = z a tolopogical argument show that|f − g | ≤ C |z |which yields the bound d < K and our second claim.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Let 1 < K = k−1k+1 <∞. Then
deg(f − g) ≤ K 2
Proof: Follows from writing f − g = P(H) P an holomorphicpolynomial and H a normalized quasiconformal mapping
|z |dK ≤ |f − g | ≤ C |z |K
Corollary if K 2 < 2, d = 1 and thus f = g .In the case, g = z a tolopogical argument show that|f − g | ≤ C |z |which yields the bound d < K and our second claim.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Let 1 < K = k−1k+1 <∞. Then
deg(f − g) ≤ K 2
Proof: Follows from writing f − g = P(H) P an holomorphicpolynomial and H a normalized quasiconformal mapping
|z |dK ≤ |f − g | ≤ C |z |K
Corollary if K 2 < 2, d = 1 and thus f = g .In the case, g = z a tolopogical argument show that|f − g | ≤ C |z |which yields the bound d < K and our second claim.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
We say that a Beltrami equation has the uniqueness property if forevery z0, z1, ω0, ω1 ∈ C there is a unique H-quasiconformal mapsuch that f (z0) = ω0 and f (z1) = ω1.If the equation has the uniqueness property we can considerFH = Φa
1 Φa(0) = 0,2 Φa(1) = a3 Lemma Φb − Φa is quasiconformal!
We say F is a family of QC mpas if 1− 3 holds.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
We say that a Beltrami equation has the uniqueness property if forevery z0, z1, ω0, ω1 ∈ C there is a unique H-quasiconformal mapsuch that f (z0) = ω0 and f (z1) = ω1.If the equation has the uniqueness property we can considerFH = Φa
1 Φa(0) = 0,2 Φa(1) = a3 Lemma Φb − Φa is quasiconformal!
We say F is a family of QC mpas if 1− 3 holds.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
We say that a Beltrami equation has the uniqueness property if forevery z0, z1, ω0, ω1 ∈ C there is a unique H-quasiconformal mapsuch that f (z0) = ω0 and f (z1) = ω1.If the equation has the uniqueness property we can considerFH = Φa
1 Φa(0) = 0,2 Φa(1) = a3 Lemma Φb − Φa is quasiconformal!
We say F is a family of QC mpas if 1− 3 holds.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
We say that a Beltrami equation has the uniqueness property if forevery z0, z1, ω0, ω1 ∈ C there is a unique H-quasiconformal mapsuch that f (z0) = ω0 and f (z1) = ω1.If the equation has the uniqueness property we can considerFH = Φa
1 Φa(0) = 0,2 Φa(1) = a3 Lemma Φb − Φa is quasiconformal!
We say F is a family of QC mpas if 1− 3 holds.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Description of FH
C linear Beltrami equation.Fµ = aΦ1, a ∈ CA complex line in W 1,2
loc (Ω,C)R linear Beltrami equationFµ,ν = aΦ1 + bΦi , a, b ∈ R(Astala-Iwaniec-Martin-Leonetti-Nesi-Jääskelainen)A two dimensional R linear subspace of R2.H linear Beltrami equation.THM ACFJ14If H(z ,w) is C 1(w)FH is an C 1-embedded submanifold of W 1,2
loc(Embedded submanifolds of Frechet spaces are inmersions suchthat the Daϕa splits)
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Description of FH
C linear Beltrami equation.Fµ = aΦ1, a ∈ CA complex line in W 1,2
loc (Ω,C)R linear Beltrami equationFµ,ν = aΦ1 + bΦi , a, b ∈ R(Astala-Iwaniec-Martin-Leonetti-Nesi-Jääskelainen)A two dimensional R linear subspace of R2.H linear Beltrami equation.THM ACFJ14If H(z ,w) is C 1(w)FH is an C 1-embedded submanifold of W 1,2
loc(Embedded submanifolds of Frechet spaces are inmersions suchthat the Daϕa splits)
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
What is the tangent plane?
THM ACFJ14 The tangent plane is given by the solutions to the Rlinear equations.
TΦaFH = Fµa,νa
where
µa(z) = ∂wH(z , ∂zφa(z)
)and νa(z) = ∂wH
(z , ∂zφa(z)
),
• ηet = φa+te−φa
t , t ∈ (0,∞), e ∈ C are QC mapsηet (0) = 0, ηe
t (1) = e normal!!• ∂zη
etj (z) == µa(z) ∂zη
etj (z) + νa(z) ∂zηe
tj (z) + htj (z),and
htj (z) =H(z ,w0 + tj ∂zη
etj (z)
)−H
(z ,w0
)tj
−∂wH
(z ,w0
)tj ∂zη
etj (z) + ∂wH
(z ,w0
)tj ∂zηe
tj (z)
tj.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
What is the tangent plane?
THM ACFJ14 The tangent plane is given by the solutions to the Rlinear equations.
TΦaFH = Fµa,νa
where
µa(z) = ∂wH(z , ∂zφa(z)
)and νa(z) = ∂wH
(z , ∂zφa(z)
),
• ηet = φa+te−φa
t , t ∈ (0,∞), e ∈ C are QC mapsηet (0) = 0, ηe
t (1) = e normal!!• ∂zη
etj (z) == µa(z) ∂zη
etj (z) + νa(z) ∂zηe
tj (z) + htj (z),and
htj (z) =H(z ,w0 + tj ∂zη
etj (z)
)−H
(z ,w0
)tj
−∂wH
(z ,w0
)tj ∂zη
etj (z) + ∂wH
(z ,w0
)tj ∂zηe
tj (z)
tj.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
What is the tangent plane?
THM ACFJ14 The tangent plane is given by the solutions to the Rlinear equations.
TΦaFH = Fµa,νa
where
µa(z) = ∂wH(z , ∂zφa(z)
)and νa(z) = ∂wH
(z , ∂zφa(z)
),
• ηet = φa+te−φa
t , t ∈ (0,∞), e ∈ C are QC mapsηet (0) = 0, ηe
t (1) = e normal!!• ∂zη
etj (z) == µa(z) ∂zη
etj (z) + νa(z) ∂zηe
tj (z) + htj (z),and
htj (z) =H(z ,w0 + tj ∂zη
etj (z)
)−H
(z ,w0
)tj
−∂wH
(z ,w0
)tj ∂zη
etj (z) + ∂wH
(z ,w0
)tj ∂zηe
tj (z)
tj.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
What is the tangent plane?
THM ACFJ14 The tangent plane is given by the solutions to the Rlinear equations.
TΦaFH = Fµa,νa
where
µa(z) = ∂wH(z , ∂zφa(z)
)and νa(z) = ∂wH
(z , ∂zφa(z)
),
• ηet = φa+te−φa
t , t ∈ (0,∞), e ∈ C are QC mapsηet (0) = 0, ηe
t (1) = e normal!!• ∂zη
etj (z) == µa(z) ∂zη
etj (z) + νa(z) ∂zηe
tj (z) + htj (z),and
htj (z) =H(z ,w0 + tj ∂zη
etj (z)
)−H
(z ,w0
)tj
−∂wH
(z ,w0
)tj ∂zη
etj (z) + ∂wH
(z ,w0
)tj ∂zηe
tj (z)
tj.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
More details
• Choosing tj sparse enough the bad set has measure zero.
B =∞⋃l=1
∞⋂k=1
∞⋃j=k
z ∈ D(0,R) : tj |∂zη
etj (z)| > 2−l
.
limj→∞
tj |∂zηetj (z)| = 0⇒
|htj (z)||∂zηe
tj (z)|→ 0, j →∞, a.e. on D(0,R).
•The limit map Φe ⊂ Fµa,νa by uniqueness the whole sequenceconverges in L∞(loc)
•Since the sequence ηet is an approximate solution to the linear
equation, approximate versions of Cacciopoli inequality yieldconvergence in the W 1,2 norm
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Questions
Theorem [AIM]"There is a one to one corrspondence between R 2 dimensional"planes" of quasiconformal mappings and linear Beltrami equations.FΦ1,Φ2 = Fµ,νTo each H we can associate a family FH.To each family F we can associate an HF if w = ∂zϕa(z) Then
H(z ,w) = ∂zϕa(z)
By quasiregularity of ϕa − ϕb this is always well defined but...howmuch this commutes?• HFH = H?• FHF = F?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Questions
Theorem [AIM]"There is a one to one corrspondence between R 2 dimensional"planes" of quasiconformal mappings and linear Beltrami equations.FΦ1,Φ2 = Fµ,νTo each H we can associate a family FH.To each family F we can associate an HF if w = ∂zϕa(z) Then
H(z ,w) = ∂zϕa(z)
By quasiregularity of ϕa − ϕb this is always well defined but...howmuch this commutes?• HFH = H?• FHF = F?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Questions
Theorem [AIM]"There is a one to one corrspondence between R 2 dimensional"planes" of quasiconformal mappings and linear Beltrami equations.FΦ1,Φ2 = Fµ,νTo each H we can associate a family FH.To each family F we can associate an HF if w = ∂zϕa(z) Then
H(z ,w) = ∂zϕa(z)
By quasiregularity of ϕa − ϕb this is always well defined but...howmuch this commutes?• HFH = H?• FHF = F?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Questions
Theorem [AIM]"There is a one to one corrspondence between R 2 dimensional"planes" of quasiconformal mappings and linear Beltrami equations.FΦ1,Φ2 = Fµ,νTo each H we can associate a family FH.To each family F we can associate an HF if w = ∂zϕa(z) Then
H(z ,w) = ∂zϕa(z)
By quasiregularity of ϕa − ϕb this is always well defined but...howmuch this commutes?• HFH = H?• FHF = F?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
answers
Key:a→ ∂aϕ(z) is a global homeomorphism in C. (for every z ∈ C)O.K under regularity assumptions of HFH does not degenerate• Local homeo in C:
det[Da∂zφa(z)] 6= 0,
(Schauder estimates+New Null Lagrangians)•∂zφ∞(z) =∞
1c≤ |∂zφa(z)|
|a|≤ c , z ∈ D(0,R), a 6= 0.
(Non vanishing of Jacobians)•Topology implies that this is enough to be a globalhomeomorphism.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 1
f (z) := ∂a1φa(z) = ∂aφa(z) + ∂aφa(z)
g(z) := ∂ai φa(z) = −i
(∂aφa(z)− ∂aφa(z)
).
Then
Im(∂z f ∂zg
)= |∂a∂zφa|2 − |∂a∂zφa|2 = det[Da∂zφa(z)],
Now f , g ∈ Fµa,νa , thus by (Astala-Jääskelainen,Leonetti-Nesi)
Im(∂z f ∂zg
)6= 0
a.e.z (Problems... the exceptional set depends on a)We need that µa(z) ∈ Cα to have conditions which hold for every zbutµa(z) = ∂wH
(z , ∂zφa(z)
)O.K for H such that H,DH are Hölder.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 1
f (z) := ∂a1φa(z) = ∂aφa(z) + ∂aφa(z)
g(z) := ∂ai φa(z) = −i
(∂aφa(z)− ∂aφa(z)
).
Then
Im(∂z f ∂zg
)= |∂a∂zφa|2 − |∂a∂zφa|2 = det[Da∂zφa(z)],
Now f , g ∈ Fµa,νa , thus by (Astala-Jääskelainen,Leonetti-Nesi)
Im(∂z f ∂zg
)6= 0
a.e.z (Problems... the exceptional set depends on a)We need that µa(z) ∈ Cα to have conditions which hold for every zbutµa(z) = ∂wH
(z , ∂zφa(z)
)O.K for H such that H,DH are Hölder.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 1
f (z) := ∂a1φa(z) = ∂aφa(z) + ∂aφa(z)
g(z) := ∂ai φa(z) = −i
(∂aφa(z)− ∂aφa(z)
).
Then
Im(∂z f ∂zg
)= |∂a∂zφa|2 − |∂a∂zφa|2 = det[Da∂zφa(z)],
Now f , g ∈ Fµa,νa , thus by (Astala-Jääskelainen,Leonetti-Nesi)
Im(∂z f ∂zg
)6= 0
a.e.z (Problems... the exceptional set depends on a)We need that µa(z) ∈ Cα to have conditions which hold for every zbutµa(z) = ∂wH
(z , ∂zφa(z)
)O.K for H such that H,DH are Hölder.
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Schauder Estimates for Non Linear Beltrami Equation
[ THM Astala-Clop-F-Jääskeläinen 14]• H(z ,w) is Cα(z), Then [Df ]Cβ < C , β ≤ minα, 1
K HereK = k+1
k−1 .Key Idea on the Proof: Perturbation from the Autonomousequation
∂z f = H(∂z f )
Then fh = f (z+h)−f (z)h is quasiregular, and thus ∂e f is quasiregular
and thus locally 1K Hölder.
Open question: Is this sharp?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Schauder Estimates for Non Linear Beltrami Equation
[ THM Astala-Clop-F-Jääskeläinen 14]• H(z ,w) is Cα(z), Then [Df ]Cβ < C , β ≤ minα, 1
K HereK = k+1
k−1 .Key Idea on the Proof: Perturbation from the Autonomousequation
∂z f = H(∂z f )
Then fh = f (z+h)−f (z)h is quasiregular, and thus ∂e f is quasiregular
and thus locally 1K Hölder.
Open question: Is this sharp?
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Schauder Estimates II
Non homogeneus, non linear Beltrami∂z f (z) = H(z , ∂z f (z)) + h(z)
Theorem (ACFJ15)
0 < α < 1 h ∈ Cαloc(C,C), let f ∈W 1,2loc (C,C) be a solution to the
non-homogeneous nonlinear Beltrami equation tw 7→ H(z ,w) ∈ C 1(C,C). Then f ∈ C 1,α
loc (C,C).
Key of the proof: Different Already at the level of the automousequation, which we can honestly differentiate
∂zg(z) = ∂wH(∂zF (z)) ∂zg(z) + ∂wH(∂zF (z)) (1)
and freeze coefficients here to get that g ∈ C 1,β for all β < 1.Then we continue with,(Campanato characterization of Hölder continuity) +Properties oflocal Beurling transform+ Cacciopoli type inequalities
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 2
• ∂zφ∞(z) =∞
1c≤ |∂zφa(z)|
|a|≤ c , z ∈ D(0,R), a 6= 0.
The map ϕa(z)a is a normalized QC map which solves a H for where
H(z ,w) = 1a H(z , aw). Thus the upper bound follows from the
Schauder estimates. Lower bound follows from
Theorem (ACFJ 14)
Let H(z ,w) is Cα(z),. Then H quasiconformal homeomorphism fsatisfy that
J(z , f ) > c
C = C (‖H‖)
C linear equation easy as ∂z f = ew , but no available,....genuinenonlinear proof
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 2
• ∂zφ∞(z) =∞
1c≤ |∂zφa(z)|
|a|≤ c , z ∈ D(0,R), a 6= 0.
The map ϕa(z)a is a normalized QC map which solves a H for where
H(z ,w) = 1a H(z , aw). Thus the upper bound follows from the
Schauder estimates. Lower bound follows from
Theorem (ACFJ 14)
Let H(z ,w) is Cα(z),. Then H quasiconformal homeomorphism fsatisfy that
J(z , f ) > c
C = C (‖H‖)
C linear equation easy as ∂z f = ew , but no available,....genuinenonlinear proof
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Non degeneracy 2
• ∂zφ∞(z) =∞
1c≤ |∂zφa(z)|
|a|≤ c , z ∈ D(0,R), a 6= 0.
The map ϕa(z)a is a normalized QC map which solves a H for where
H(z ,w) = 1a H(z , aw). Thus the upper bound follows from the
Schauder estimates. Lower bound follows from
Theorem (ACFJ 14)
Let H(z ,w) is Cα(z),. Then H quasiconformal homeomorphism fsatisfy that
J(z , f ) > c
C = C (‖H‖)
C linear equation easy as ∂z f = ew , but no available,....genuinenonlinear proof
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Proof of Non vanishing of Jacobians
Autonomous equation • fh = f (z+h)−f (z)h is quasiregular and never
vanish because f is homeo.•Hurwitz type theorem implies that the limit does not vanish. UseMiniowitz or factorization fh = hh Fh and compactness for Fh +Hurwitz type theorem for hh. General equation.• Perturbation as in the proof of Schauder estimates impliespositivity of J(z , f )• Compactness argument gives the lower bound
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
final theorems
Theorem
Suppose that H is a regular field. Then FH defines H uniquely,that is, HFH = H.
Theorem
F = φa(z)a∈C non-degenerate family and regular, i.e
‖Dzφa‖C s(DR) ≤ c(s,R) |a|, R <∞, (2)
and DaDzF ∈ (C (z),C (a)) Then there is a unique H = HF ,H(z ,w) ∈ C s
loc(z) , DwH(z ,w) ∈ (C (z),C (w)).
TheoremUnder ellipticity bounds at ∞ FHF = F .
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
final theorems
Theorem
Suppose that H is a regular field. Then FH defines H uniquely,that is, HFH = H.
Theorem
F = φa(z)a∈C non-degenerate family and regular, i.e
‖Dzφa‖C s(DR) ≤ c(s,R) |a|, R <∞, (2)
and DaDzF ∈ (C (z),C (a)) Then there is a unique H = HF ,H(z ,w) ∈ C s
loc(z) , DwH(z ,w) ∈ (C (z),C (w)).
TheoremUnder ellipticity bounds at ∞ FHF = F .
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
final theorems
Theorem
Suppose that H is a regular field. Then FH defines H uniquely,that is, HFH = H.
Theorem
F = φa(z)a∈C non-degenerate family and regular, i.e
‖Dzφa‖C s(DR) ≤ c(s,R) |a|, R <∞, (2)
and DaDzF ∈ (C (z),C (a)) Then there is a unique H = HF ,H(z ,w) ∈ C s
loc(z) , DwH(z ,w) ∈ (C (z),C (w)).
TheoremUnder ellipticity bounds at ∞ FHF = F .
D.Faraco (UAM-ICMAT) Non linear Beltrami equation
Questions
• What are the advantage of the manifold interpretation? Is there aFrobenius theorem? Can we propagate properties of solutions tothe linear equation to the nonlinear through the exponential map?What is the relation between the curvature of H and that of FH?• How much of the regularity is needed in the last theorem. Inprinciple the exceptional sets depend on a...? Lots of measurabilityproblems....
D.Faraco (UAM-ICMAT) Non linear Beltrami equation