En esta charla mostramos que la extensio ́n de un operador definido en un espacio de funciones de Banach es equivalente, bajo ciertos requisitos m ́ınimos, a su factorizacio ́natrave ́sdeunsubespaciodefuncionesintegrablesrespectodela medidavectorialmT asociadaaloperadorT.Adema ́s,estesubespacioparticular determina aquella propiedad del operador que se pretende preservar, en el sentido de queconstituyeeldominioo ́ptimo-elmayorsubespacioalquesepuedeextender-,de manera que la extensio ́n conserva esta propiedad. Esto permite un planteamiento abstracto de ciertos aspectos de la teor ́ıa de dominio o ́ptimo, que se ha desarrollado de manera importante en los u ́ltimos tiempos, principalmente en la direccio ́n de caracterizar cua ́l es el mayor espacio de funciones de Banach al que se pueden extender algunos operadores de importancia en ana ́lisis. Adema ́s, permite clasificar laspropiedadesdelosoperadoresqueseconservanenlaextensio ́nenfuncio ́n delsubespaciodeL1(mT)queconstituyesudominioo ́ptimo.
Subespacios de funciones integrables respecto de una medidavectorialyextensio ́ndeoperadores
1. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Subespacios de funciones integrables respecto de una
medida vectorial y extensi´on de operadores
Enrique A. S´anchez P´erez
J. M. Calabuig O. Delgado
Instituto de Matem´atica Pura y Aplicada (I.M.P.A.).
Departamento de Matem´atica Aplicada.
Universidad Polit´ecnica de Valencia.
Camino de Vera S/N, 46022, Valencia.
X EARCO
Mallorca, Mayo 2007
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
2. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
En esta charla mostramos que la extensi´on de un operador definido en un espacio de
funciones de Banach es equivalente, bajo ciertos requisitos m´ınimos, a su
factorizaci´on a trav´es de un subespacio de funciones integrables respecto de la
medida vectorial mT asociada al operador T. Adem´as, este subespacio particular
determina aquella propiedad del operador que se pretende preservar, en el sentido de
que constituye el dominio ´optimo -el mayor subespacio al que se puede extender-, de
manera que la extensi´on conserva esta propiedad. Esto permite un planteamiento
abstracto de ciertos aspectos de la teor´ıa de dominio ´optimo, que se ha desarrollado
de manera importante en los ´ultimos tiempos, principalmente en la direcci´on de
caracterizar cu´al es el mayor espacio de funciones de Banach al que se pueden
extender algunos operadores de importancia en an´alisis. Adem´as, permite clasificar
las propiedades de los operadores que se conservan en la extensi´on en funci´on
del subespacio de L1(mT ) que constituye su dominio ´optimo.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
3. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Index of the talk
1 Notation and preliminaries
2 Vector measures and spaces of integrable functions
3 A factorization theorem
4 An application: extensions of operators on p-convex B. f. s.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
4. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Index of the talk
1 Notation and preliminaries
2 Vector measures and spaces of integrable functions
3 A factorization theorem
4 An application: extensions of operators on p-convex B. f. s.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
5. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Index of the talk
1 Notation and preliminaries
2 Vector measures and spaces of integrable functions
3 A factorization theorem
4 An application: extensions of operators on p-convex B. f. s.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
6. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Index of the talk
1 Notation and preliminaries
2 Vector measures and spaces of integrable functions
3 A factorization theorem
4 An application: extensions of operators on p-convex B. f. s.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
7. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
FIRST PART
Notation and preliminaries
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
8. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
General notation
(Ω,Σ,µ) will be a finite measure space,
E be a Banach space and E its topological dual,
BE the unit ball of E,
P(A) will represent the set of partitions π of A into finite number of disjoint
measurable sets,
m : Σ → E will be a (countably additive) vector measure,
The semivariation of m over A ∈ Σ is defined by
m (A) = sup
x ∈BE
| m,x |(A) = sup
x ∈BE
sup
π∈P(A)
∑
B∈π
| m,x (B)|,
with the usual notation
m,x (B) = m(B),x for each B ∈ Σ,
A ∈ Σ is called m-null if m (A) = 0.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
9. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
General notation
(Ω,Σ,µ) will be a finite measure space,
E be a Banach space and E its topological dual,
BE the unit ball of E,
P(A) will represent the set of partitions π of A into finite number of disjoint
measurable sets,
m : Σ → E will be a (countably additive) vector measure,
The semivariation of m over A ∈ Σ is defined by
m (A) = sup
x ∈BE
| m,x |(A) = sup
x ∈BE
sup
π∈P(A)
∑
B∈π
| m,x (B)|,
with the usual notation
m,x (B) = m(B),x for each B ∈ Σ,
A ∈ Σ is called m-null if m (A) = 0.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
10. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Remark
In this talk we assume that µ and m are mutually absolutely continuous.
In particular, µ can be a Rybakov measure; recall that a Rybakov measure for a vector
measure m is a scalar measure ν defined as the variation of a measure m,x , where
x ∈ E , whenever m is absolutely continuous with respect to ν.
A Rybakov measure always exists for every vector measure m (see [5, IX.2.2]).
Banach function (sub)spaces.
Let L0(µ) be the space of all measurable real functions on Ω.
1 We will call Banach function space to any Banach space X(µ) ⊆ L0(µ) with norm
· X(µ) satisfying that
1) if f ∈ L0(µ), g ∈ X(µ) with |f| ≤ |g| µ-a.e. then f ∈ X(µ) and f X(µ) ≤ g X(µ).
2) For every A ∈ Σ, the characteristic function χA belongs to X(µ).
2 By a Banach function subspace of X(µ) we mean a Banach function space
continuously contained in X(µ), with the same order structure (allowing different
norms).
3 A Banach function space is order continuous if order bounded increasing
sequences are convergent in norm.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
11. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Remark
In this talk we assume that µ and m are mutually absolutely continuous.
In particular, µ can be a Rybakov measure; recall that a Rybakov measure for a vector
measure m is a scalar measure ν defined as the variation of a measure m,x , where
x ∈ E , whenever m is absolutely continuous with respect to ν.
A Rybakov measure always exists for every vector measure m (see [5, IX.2.2]).
Banach function (sub)spaces.
Let L0(µ) be the space of all measurable real functions on Ω.
1 We will call Banach function space to any Banach space X(µ) ⊆ L0(µ) with norm
· X(µ) satisfying that
1) if f ∈ L0(µ), g ∈ X(µ) with |f| ≤ |g| µ-a.e. then f ∈ X(µ) and f X(µ) ≤ g X(µ).
2) For every A ∈ Σ, the characteristic function χA belongs to X(µ).
2 By a Banach function subspace of X(µ) we mean a Banach function space
continuously contained in X(µ), with the same order structure (allowing different
norms).
3 A Banach function space is order continuous if order bounded increasing
sequences are convergent in norm.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
12. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
L1(m)L1(m)L1(m) spaces.
A function f : Ω → R is said to be integrable with respect to the measure m if
1 for each x ∈ E we have that f ∈ L1( m,x ),
2 for each A ∈ Σ there exists xA ∈ E such that
xA,x =
A
fd m,x for every x ∈ E .
The element xA is usually denoted by A f dm. The space of the classes (equality
m-almost everywhere) of these functions is denoted by L1(m). The expression
f m = sup
x ∈BE
|f|d| m,x | for each f ∈ L1
(m),
defines a lattice norm on L1(m) for which L1(m) is an order continuous Banach
function space.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
13. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
The indefinite integral mf : Σ → E of a function f ∈ L1(m) is defined by
mf (A) =
A
fdm, A ∈ Σ.
The Orlicz-Pettis Theorem ensures that mf is again a countably additive vector
measure. An equivalent norm for L1(m) is given by
|||f|||m = sup
A∈Σ A
fdm
E
for each f ∈ L1
(m).
Moreover |||f|||L1(m) ≤ f L1(m) ≤ 2|||f|||L1(m) ([5, Ch. I.1.11]).
Lp(m)Lp(m)Lp(m) spaces.
Let 1 ≤ p < ∞. A measurable function f : Ω → R is said to be p-integrable if |f|p is
integrable with respect to m. The expression.
f Lp(m) = sup
x ∈BE
|f|p
d| m,x |
1
p
for each f ∈ Lp
(m),
defines a lattice norm on Lp(m) ([6]). The space Lp(m) is an order continuous Banach
function subspace of L1(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
14. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
The indefinite integral mf : Σ → E of a function f ∈ L1(m) is defined by
mf (A) =
A
fdm, A ∈ Σ.
The Orlicz-Pettis Theorem ensures that mf is again a countably additive vector
measure. An equivalent norm for L1(m) is given by
|||f|||m = sup
A∈Σ A
fdm
E
for each f ∈ L1
(m).
Moreover |||f|||L1(m) ≤ f L1(m) ≤ 2|||f|||L1(m) ([5, Ch. I.1.11]).
Lp(m)Lp(m)Lp(m) spaces.
Let 1 ≤ p < ∞. A measurable function f : Ω → R is said to be p-integrable if |f|p is
integrable with respect to m. The expression.
f Lp(m) = sup
x ∈BE
|f|p
d| m,x |
1
p
for each f ∈ Lp
(m),
defines a lattice norm on Lp(m) ([6]). The space Lp(m) is an order continuous Banach
function subspace of L1(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
15. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
The indefinite integral mf : Σ → E of a function f ∈ L1(m) is defined by
mf (A) =
A
fdm, A ∈ Σ.
The Orlicz-Pettis Theorem ensures that mf is again a countably additive vector
measure. An equivalent norm for L1(m) is given by
|||f|||m = sup
A∈Σ A
fdm
E
for each f ∈ L1
(m).
Moreover |||f|||L1(m) ≤ f L1(m) ≤ 2|||f|||L1(m) ([5, Ch. I.1.11]).
Lp(m)Lp(m)Lp(m) spaces.
Let 1 ≤ p < ∞. A measurable function f : Ω → R is said to be p-integrable if |f|p is
integrable with respect to m. The expression.
f Lp(m) = sup
x ∈BE
|f|p
d| m,x |
1
p
for each f ∈ Lp
(m),
defines a lattice norm on Lp(m) ([6]). The space Lp(m) is an order continuous Banach
function subspace of L1(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
16. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
SECOND PART
Vector measures and spaces of integrable functions.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
17. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Y(µ)Y(µ)Y(µ) semivariation.
Let m : Σ → E be a (countably additive) vector measure and Y(µ) a Banach function
space. The Y(µ)-semivariation m Y(µ) of m is defined by
m Y(µ) := sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BY(µ) .
L1
Y(µ)(m)L1
Y(µ)(m)L1
Y(µ)(m) spaces.
We define L1
Y(µ)(m) as the space of all (classes of m-a.e. equal) functions f of L1(m)
such that the vector measure associated to m, mf , has finite Y(µ)-semivariation
endowed with the norm
f L1
Y(µ)
(m) := mf Y(µ) = sup
ϕ simple
ϕ∈BY(µ)
fϕdm
E
.
L1
Y(µ)(m) is always a Banach function subspace of L1(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
18. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Y(µ)Y(µ)Y(µ) semivariation.
Let m : Σ → E be a (countably additive) vector measure and Y(µ) a Banach function
space. The Y(µ)-semivariation m Y(µ) of m is defined by
m Y(µ) := sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BY(µ) .
L1
Y(µ)(m)L1
Y(µ)(m)L1
Y(µ)(m) spaces.
We define L1
Y(µ)(m) as the space of all (classes of m-a.e. equal) functions f of L1(m)
such that the vector measure associated to m, mf , has finite Y(µ)-semivariation
endowed with the norm
f L1
Y(µ)
(m) := mf Y(µ) = sup
ϕ simple
ϕ∈BY(µ)
fϕdm
E
.
L1
Y(µ)(m) is always a Banach function subspace of L1(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
19. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Some particular subspaces of L1(m) that are nowadays well known can be
represented in terms of the spaces L1
Y(µ)(m).
1 The semivariation of a vector measure can be computed by
m = sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BL∞(µ) = m L∞(µ),
where µ is a finite positive measure equivalent to m (see [5, Ch.I]). Thus, for any
function f ∈ L1(m), mf = mf L∞(µ) = f L1
L∞(µ)
(m).
Hence, L1(m) = L1
L∞(µ)(m).
2 Let 1 < p < ∞ and 1/p +1/p = 1. It is known that for any function f ∈ Lp(m), the
expression
mf Lp (m)
= sup
ϕ simple
ϕ∈B
Lp (m)
fϕdm
E
equals the norm f L
p
w (m)
of a function f belonging to L
p
w (m) ( [7] ).
In fact, it can be proved that L
p
w (m) = L1
Lp (m)
(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
20. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Some particular subspaces of L1(m) that are nowadays well known can be
represented in terms of the spaces L1
Y(µ)(m).
1 The semivariation of a vector measure can be computed by
m = sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BL∞(µ) = m L∞(µ),
where µ is a finite positive measure equivalent to m (see [5, Ch.I]). Thus, for any
function f ∈ L1(m), mf = mf L∞(µ) = f L1
L∞(µ)
(m).
Hence, L1(m) = L1
L∞(µ)(m).
2 Let 1 < p < ∞ and 1/p +1/p = 1. It is known that for any function f ∈ Lp(m), the
expression
mf Lp (m)
= sup
ϕ simple
ϕ∈B
Lp (m)
fϕdm
E
equals the norm f L
p
w (m)
of a function f belonging to L
p
w (m) ( [7] ).
In fact, it can be proved that L
p
w (m) = L1
Lp (m)
(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
21. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Some particular subspaces of L1(m) that are nowadays well known can be
represented in terms of the spaces L1
Y(µ)(m).
1 The semivariation of a vector measure can be computed by
m = sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BL∞(µ) = m L∞(µ),
where µ is a finite positive measure equivalent to m (see [5, Ch.I]). Thus, for any
function f ∈ L1(m), mf = mf L∞(µ) = f L1
L∞(µ)
(m).
Hence, L1(m) = L1
L∞(µ)(m).
2 Let 1 < p < ∞ and 1/p +1/p = 1. It is known that for any function f ∈ Lp(m), the
expression
mf Lp (m)
= sup
ϕ simple
ϕ∈B
Lp (m)
fϕdm
E
equals the norm f L
p
w (m)
of a function f belonging to L
p
w (m) ( [7] ).
In fact, it can be proved that L
p
w (m) = L1
Lp (m)
(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
22. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Some particular subspaces of L1(m) that are nowadays well known can be
represented in terms of the spaces L1
Y(µ)(m).
1 The semivariation of a vector measure can be computed by
m = sup
n
∑
i=1
αAi
m(Ai )
E
:
n
∑
i=1
αAi
χAi
∈ BL∞(µ) = m L∞(µ),
where µ is a finite positive measure equivalent to m (see [5, Ch.I]). Thus, for any
function f ∈ L1(m), mf = mf L∞(µ) = f L1
L∞(µ)
(m).
Hence, L1(m) = L1
L∞(µ)(m).
2 Let 1 < p < ∞ and 1/p +1/p = 1. It is known that for any function f ∈ Lp(m), the
expression
mf Lp (m)
= sup
ϕ simple
ϕ∈B
Lp (m)
fϕdm
E
equals the norm f L
p
w (m)
of a function f belonging to L
p
w (m) ( [7] ).
In fact, it can be proved that L
p
w (m) = L1
Lp (m)
(m).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
23. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
THIRD PART
A factorization theorem.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
24. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Let X(µ) be an order continuous Banach function space and let T : X(µ) → E be an
operator. As usual, we define the (countably additive) vector measure
mT : Σ → E by mT (A) := T(χA).
An operator T is µ-determined if the set of mT -null sets equals the set of µ-null sets.
Let Y(µ) be a Banach function space over µ containing the set of the simple functions.
Y(µ)Y(µ)Y(µ)-extensible operators.
A µ-determined operator T : X(µ) → E is said to be Y(µ)-extensible if there is a
constant K > 0 such that
T(fϕ) ≤ K f X(µ) ϕ Y(µ)
for every function f ∈ X(µ) and every simple function ϕ.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
25. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Let X(µ) be an order continuous Banach function space and let T : X(µ) → E be an
operator. As usual, we define the (countably additive) vector measure
mT : Σ → E by mT (A) := T(χA).
An operator T is µ-determined if the set of mT -null sets equals the set of µ-null sets.
Let Y(µ) be a Banach function space over µ containing the set of the simple functions.
Y(µ)Y(µ)Y(µ)-extensible operators.
A µ-determined operator T : X(µ) → E is said to be Y(µ)-extensible if there is a
constant K > 0 such that
T(fϕ) ≤ K f X(µ) ϕ Y(µ)
for every function f ∈ X(µ) and every simple function ϕ.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
26. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Let X(µ) be an order continuous Banach function space and let T : X(µ) → E be an
operator. As usual, we define the (countably additive) vector measure
mT : Σ → E by mT (A) := T(χA).
An operator T is µ-determined if the set of mT -null sets equals the set of µ-null sets.
Let Y(µ) be a Banach function space over µ containing the set of the simple functions.
Y(µ)Y(µ)Y(µ)-extensible operators.
A µ-determined operator T : X(µ) → E is said to be Y(µ)-extensible if there is a
constant K > 0 such that
T(fϕ) ≤ K f X(µ) ϕ Y(µ)
for every function f ∈ X(µ) and every simple function ϕ.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
27. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Theorem (First part)
Let X(µ) be an order continuous Banach function space with and let Y(µ) be a
Banach function space verifying that the simple functions are dense. Given a
µ-determined operator T : X(µ) → E the following assertions are equivalent
1 T is Y(µ)-extensible.
2 The bounded linear map T can be factored through the space L1
Y(µ)(mT ) as
X(µ)
T
//
i
$$
E
L1
Y(µ)(mT )
IY(µ)
;;
where the integration map IY(µ) is also Y(µ)-extensible.
3 There is a constant K > 0 verifying that for every f ∈ L1(m)
(mT )f Y(µ) ≤ K f X(µ)
where (mT )f : Σ → E is the vector measure given by (mT )f (A) = A fdmT .
4 X(µ)·Y(µ) ⊆ L1(mT ).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
28. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Theorem (First part)
Let X(µ) be an order continuous Banach function space with and let Y(µ) be a
Banach function space verifying that the simple functions are dense. Given a
µ-determined operator T : X(µ) → E the following assertions are equivalent
1 T is Y(µ)-extensible.
2 The bounded linear map T can be factored through the space L1
Y(µ)(mT ) as
X(µ)
T
//
i
$$
E
L1
Y(µ)(mT )
IY(µ)
;;
where the integration map IY(µ) is also Y(µ)-extensible.
3 There is a constant K > 0 verifying that for every f ∈ L1(m)
(mT )f Y(µ) ≤ K f X(µ)
where (mT )f : Σ → E is the vector measure given by (mT )f (A) = A fdmT .
4 X(µ)·Y(µ) ⊆ L1(mT ).
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
29. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Theorem (Second part, optimal domain)
Moreover, if Z(µ) is any other order continuous Banach function space such that
X(µ) ⊆ Z(µ) with continuous inclusion,
T can be extended to Z(µ) and this extension is Y(µ)-extensible,
then Z(µ) ⊆ L1
Y(µ)(mT ) with continuous inclusion.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
30. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Assume that T is continuous. In this case taking into account that for each f ∈ X(µ)
and each ϕ simple function
T(fϕ) E ≤ K fϕ X(µ) ≤ K f X(µ) ϕ L∞(µ),
we obtain that T is L∞(µ)-extensible. Reciprocally if T is L∞(µ)-extensible then taking
ϕ = χΩ we obtain that T is continuous. Applying Theorem 3 we obtain that the optimal
domain is L1
L∞(µ)(m) = L1(m) so this result includes the case studied by Curbera G.P.
and Ricker W. in [2].
Example ([1])
In [1] the authors introduce the notion of Lp-product extensible operator and define the
spaces L1
p,µ (m). In this article they prove that the bigger function space (i.e. the optimal
domain) to which an Lp-product extensible operator T can be extended preserving this
property is exactly the space L1
p,µ (mT ). Actually the notion of Lp-product extensible
operator coincides with our definition of Y(µ)-extensible operator with Y(µ) = Lp(µ) for
1 < p < ∞. In this case the space L1
p,µ (mT ) is now L1
Lp(µ)(mT ). Hence Theorem 6 in [1]
is a particular case of Theorem 3.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
31. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Example
Assume that T is continuous. In this case taking into account that for each f ∈ X(µ)
and each ϕ simple function
T(fϕ) E ≤ K fϕ X(µ) ≤ K f X(µ) ϕ L∞(µ),
we obtain that T is L∞(µ)-extensible. Reciprocally if T is L∞(µ)-extensible then taking
ϕ = χΩ we obtain that T is continuous. Applying Theorem 3 we obtain that the optimal
domain is L1
L∞(µ)(m) = L1(m) so this result includes the case studied by Curbera G.P.
and Ricker W. in [2].
Example ([1])
In [1] the authors introduce the notion of Lp-product extensible operator and define the
spaces L1
p,µ (m). In this article they prove that the bigger function space (i.e. the optimal
domain) to which an Lp-product extensible operator T can be extended preserving this
property is exactly the space L1
p,µ (mT ). Actually the notion of Lp-product extensible
operator coincides with our definition of Y(µ)-extensible operator with Y(µ) = Lp(µ) for
1 < p < ∞. In this case the space L1
p,µ (mT ) is now L1
Lp(µ)(mT ). Hence Theorem 6 in [1]
is a particular case of Theorem 3.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
32. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
FOURTH PART
An application: extension of operators on p-convex B. f. s.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
33. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
A Banach lattice E is p-convex if there is a constant K such that for each finite set
x1,...,xn ∈ E,
(
n
∑
i=1
|xi |p
)1/p
≤ K(
n
∑
i=1
xi
p
)1/p
.
Let T : E → F be an operator, where F is a lattice. T is p-concave if there is a
constant K such that for each finite family x1,...,xn ∈ E,
(
n
∑
i=1
T(xi ) p
)1/p
≤ K (
n
∑
i=1
|xi |p
)1/p
.
If X(µ) a Banach function space we define its p-th power X(µ)[p] as the space
X(µ)[p] := {|f|p
: f ∈ X(µ)}.
If X(µ) is p-convex, X(µ)[p] is a Banach function space over µ with the quasi norm
g X(µ)[p]
:= |g|1/p p
X(µ)
, g ∈ X(µ)[p],
that in this case is equivalent to a norm.
If µ is a finite measure, then the inclusion X(µ) ⊆ X(µ)[p] is well-defined and
continuous.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
34. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
A Banach lattice E is p-convex if there is a constant K such that for each finite set
x1,...,xn ∈ E,
(
n
∑
i=1
|xi |p
)1/p
≤ K(
n
∑
i=1
xi
p
)1/p
.
Let T : E → F be an operator, where F is a lattice. T is p-concave if there is a
constant K such that for each finite family x1,...,xn ∈ E,
(
n
∑
i=1
T(xi ) p
)1/p
≤ K (
n
∑
i=1
|xi |p
)1/p
.
If X(µ) a Banach function space we define its p-th power X(µ)[p] as the space
X(µ)[p] := {|f|p
: f ∈ X(µ)}.
If X(µ) is p-convex, X(µ)[p] is a Banach function space over µ with the quasi norm
g X(µ)[p]
:= |g|1/p p
X(µ)
, g ∈ X(µ)[p],
that in this case is equivalent to a norm.
If µ is a finite measure, then the inclusion X(µ) ⊆ X(µ)[p] is well-defined and
continuous.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
35. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
A Banach lattice E is p-convex if there is a constant K such that for each finite set
x1,...,xn ∈ E,
(
n
∑
i=1
|xi |p
)1/p
≤ K(
n
∑
i=1
xi
p
)1/p
.
Let T : E → F be an operator, where F is a lattice. T is p-concave if there is a
constant K such that for each finite family x1,...,xn ∈ E,
(
n
∑
i=1
T(xi ) p
)1/p
≤ K (
n
∑
i=1
|xi |p
)1/p
.
If X(µ) a Banach function space we define its p-th power X(µ)[p] as the space
X(µ)[p] := {|f|p
: f ∈ X(µ)}.
If X(µ) is p-convex, X(µ)[p] is a Banach function space over µ with the quasi norm
g X(µ)[p]
:= |g|1/p p
X(µ)
, g ∈ X(µ)[p],
that in this case is equivalent to a norm.
If µ is a finite measure, then the inclusion X(µ) ⊆ X(µ)[p] is well-defined and
continuous.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
36. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
A Banach lattice E is p-convex if there is a constant K such that for each finite set
x1,...,xn ∈ E,
(
n
∑
i=1
|xi |p
)1/p
≤ K(
n
∑
i=1
xi
p
)1/p
.
Let T : E → F be an operator, where F is a lattice. T is p-concave if there is a
constant K such that for each finite family x1,...,xn ∈ E,
(
n
∑
i=1
T(xi ) p
)1/p
≤ K (
n
∑
i=1
|xi |p
)1/p
.
If X(µ) a Banach function space we define its p-th power X(µ)[p] as the space
X(µ)[p] := {|f|p
: f ∈ X(µ)}.
If X(µ) is p-convex, X(µ)[p] is a Banach function space over µ with the quasi norm
g X(µ)[p]
:= |g|1/p p
X(µ)
, g ∈ X(µ)[p],
that in this case is equivalent to a norm.
If µ is a finite measure, then the inclusion X(µ) ⊆ X(µ)[p] is well-defined and
continuous.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
37. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Let X(µ) be an order continuous p-convex Banach function space and E a
p-concave Banach lattice. Consider a positive operator T : X(µ) → E. Then it is
known that there is a function h belonging to the dual of X(µ)[p] such that
T(f) ≤ K |f|p
hdµ
1/p
, f ∈ X(µ).
We can assume also that h > 0 µ-a.e.
We can use this inequality to generate a norm function by
q∗
(ϕ) := sup{ T(fϕ) : |f|p
hdµ ≤ 1}, ϕ simple.
By using a density argument we can generate a function space G(µ) satisfying
that
fϕ dmT ≤ Kq∗
(ϕ) |f|p
hdµ
1/p
.
Now we can apply the factorization theorem to this inequality. We obtain in this
way an extension of T to the subspace L1
G(µ)(mT ) that is optimal in the sense that
L1
G(µ)(mT ) is the biggest space to which T can be extended still satisfying the
inequality.
The subspace L1
G(µ)(mT ) is p-convex.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores
38. Notation and preliminaries
Vector measures and spaces of integrable functions
A factorization theorem
An application: extensions of operators on p-convex B. f. s.
Some references
Calabuig J.M., Galaz F., Jim´enez Fern´andez E. and S´anchez P´erez E.A., Strong factorization
of operators on spaces of vector measure integrable functions and unconditional convergence
of series. To appear in Mathematische Zeitschrift.
Curbera, G.P. and Ricker, W.J.: Optimal domains for the kernel operator associated with
Sobolev’s inequality , Studia Math., 158(2), (2003), 131-152 (see also Corrigenda in the same
journal , 170, (2005), 217-218).
Delgado, O.: Optimal domains for kernel operators on [0,∞)×[0,∞), Studia Math. 174 (2006),
131-145.
Delgado, O. and Soria, J.: Optimal domain for the Hardy operator, J. Funct. Anal. 244 (2007),
119-133.
Diestel J. and Uhl J.J., Vector Measures, Math. Surveys, vol. 15, Amer. Math. Soc.,
Providence, RI, 1977.
Fern´andez, A., Mayoral, F., Naranjo, F., S´aez, C. and S´anchez-P´erez, E.A.: Spaces of
p-integrable functions with respect to a vector measure., Positivity, 10(2006), 1-16.
S´anchez P´erez E.A., Compactness arguments for spaces of p-integrable functions with
respect to a vector measure and factorization of operators through Lebesgue-Bochner
spaces. Illinois J. Math. 45,3(2001), 907-923.
Enrique A. S´anchez P´erez Subespacios y extensiones de operadores