spaces on simple closed rectifiable curves in the complex plane and - - PowerPoint PPT Presentation

spaces on simple closed rectifiable curves in the complex
SMART_READER_LITE
LIVE PREVIEW

spaces on simple closed rectifiable curves in the complex plane and - - PowerPoint PPT Presentation

The closed span of an exponential system in L p spaces on simple closed rectifiable curves in the complex plane and P olya singularity results for Taylor-Dirichlet series Elias Zikkos Cyprus Ministry of Education and Culture New


slide-1
SLIDE 1

The closed span of an exponential system in Lp spaces on simple closed rectifiable curves in the complex plane and P´

  • lya singularity results for

Taylor-Dirichlet series

Elias Zikkos Cyprus Ministry of Education and Culture New Developments in Complex Analysis and Function Theory Crete 2018, July 2

  • E. Zikkos

Short version

slide-2
SLIDE 2

The Polya Theorem for Exponential Dirichlet Series

  • E. Zikkos

Short version

slide-3
SLIDE 3

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers,

  • E. Zikkos

Short version

slide-4
SLIDE 4

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞.

  • E. Zikkos

Short version

slide-5
SLIDE 5

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R.

  • E. Zikkos

Short version

slide-6
SLIDE 6

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ.

  • E. Zikkos

Short version

slide-7
SLIDE 7

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute).

  • E. Zikkos

Short version

slide-8
SLIDE 8

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA:

  • E. Zikkos

Short version

slide-9
SLIDE 9

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity

  • E. Zikkos

Short version

slide-10
SLIDE 10

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity in every

  • pen interval whose length Exceeds 2πd and lies on the abscissa of

convergence.

  • E. Zikkos

Short version

slide-11
SLIDE 11

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity in every

  • pen interval whose length Exceeds 2πd and lies on the abscissa of

convergence. Example (trivial): ez 1 − ez

  • E. Zikkos

Short version

slide-12
SLIDE 12

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity in every

  • pen interval whose length Exceeds 2πd and lies on the abscissa of

convergence. Example (trivial): ez 1 − ez =

  • n=1

enz, ℜz < 0,

  • E. Zikkos

Short version

slide-13
SLIDE 13

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity in every

  • pen interval whose length Exceeds 2πd and lies on the abscissa of

convergence. Example (trivial): ez 1 − ez =

  • n=1

enz, ℜz < 0, Density = 1.

  • E. Zikkos

Short version

slide-14
SLIDE 14

The Polya Theorem for Exponential Dirichlet Series

Given a strictly increasing sequence Λ = {λn}∞

n=1, of positive real

numbers, uniformly separated and having Density d, λn+1 − λn > c > 0, lim

n→∞

n λn = d < ∞. consider the class of Dirichlet series of the form

  • n=1

cneλnz, lim sup

n→∞

log |cn| λn = ξ ∈ R. The series is analytic in the left half-plane ℜz < −ξ, it converges uniformly on compact subsets, and it diverges for all z such ℜz > −ξ. The line ℜz = −ξ is called the Abscissa of Convergence (pointwise and absolute). POLYA: the series has at least One singularity in every

  • pen interval whose length Exceeds 2πd and lies on the abscissa of

convergence. Example (trivial): ez 1 − ez =

  • n=1

enz, ℜz < 0, Density = 1. Singularities at the points 2kπi, k ∈ Z

  • E. Zikkos

Short version

slide-15
SLIDE 15

First Goal: Generalizing The Polya Theorem

  • E. Zikkos

Short version

slide-16
SLIDE 16

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . }

  • E. Zikkos

Short version

slide-17
SLIDE 17

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity,

  • E. Zikkos

Short version

slide-18
SLIDE 18

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded.

  • E. Zikkos

Short version

slide-19
SLIDE 19

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded. We impose two conditions:

  • E. Zikkos

Short version

slide-20
SLIDE 20

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded. We impose two conditions: (A) Λ has Density d counting multiplicities

  • E. Zikkos

Short version

slide-21
SLIDE 21

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded. We impose two conditions: (A) Λ has Density d counting multiplicities lim

t→∞

nΛ(t) t = d < ∞, nΛ(t) =

  • λn≤t

µn

  • E. Zikkos

Short version

slide-22
SLIDE 22

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded. We impose two conditions: (A) Λ has Density d counting multiplicities lim

t→∞

nΛ(t) t = d < ∞, nΛ(t) =

  • λn≤t

µn (if µn = 1 for all n ∈ N then n/λn → d)

  • E. Zikkos

Short version

slide-23
SLIDE 23

First Goal: Generalizing The Polya Theorem

We consider Taylor-Dirichlet series

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz

associated to a multiplicity sequence Λ = {λn, µn}∞

n=1

{λn, µn}∞

n=1 := {λ1, λ1, . . . , λ1

  • µ1−times

, λ2, λ2, . . . , λ2

  • µ2−times

, . . . , λk, λk, . . . , λk

  • µk−times

, . . . } {λn}∞

n=1 is a strictly increasing sequence of positive real numbers

diverging to infinity, AND {µn}∞

n=1 is a sequence of positive integers,

Not Necessarily Bounded. We impose two conditions: (A) Λ has Density d counting multiplicities lim

t→∞

nΛ(t) t = d < ∞, nΛ(t) =

  • λn≤t

µn (if µn = 1 for all n ∈ N then n/λn → d) (B) λn+1 − λn > c > 0, (Uniformly Separated).

  • E. Zikkos

Short version

slide-24
SLIDE 24

Region of Convergence, Taylor-Dirichlet series:

  • E. Zikkos

Short version

slide-25
SLIDE 25

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0.

  • E. Zikkos

Short version

slide-26
SLIDE 26

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) :

  • E. Zikkos

Short version

slide-27
SLIDE 27

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) : consider the series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

Cn = max{|cn,k| : k = 0, 1, 2, . . . , µn−1} lim sup

n→∞

log Cn λn = ξ ∈ R, P−ξ := {z : ℜz < −ξ}.

  • E. Zikkos

Short version

slide-28
SLIDE 28

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) : consider the series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

Cn = max{|cn,k| : k = 0, 1, 2, . . . , µn−1} lim sup

n→∞

log Cn λn = ξ ∈ R, P−ξ := {z : ℜz < −ξ}. Then g(z) is an analytic function in the left half-plane P−ξ converging uniformly on compact subsets.

  • E. Zikkos

Short version

slide-29
SLIDE 29

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) : consider the series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

Cn = max{|cn,k| : k = 0, 1, 2, . . . , µn−1} lim sup

n→∞

log Cn λn = ξ ∈ R, P−ξ := {z : ℜz < −ξ}. Then g(z) is an analytic function in the left half-plane P−ξ converging uniformly on compact subsets. We call the line ℜz = −ξ the abscissa of convergence for g(z).

  • E. Zikkos

Short version

slide-30
SLIDE 30

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) : consider the series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

Cn = max{|cn,k| : k = 0, 1, 2, . . . , µn−1} lim sup

n→∞

log Cn λn = ξ ∈ R, P−ξ := {z : ℜz < −ξ}. Then g(z) is an analytic function in the left half-plane P−ξ converging uniformly on compact subsets. We call the line ℜz = −ξ the abscissa of convergence for g(z). Question :

  • E. Zikkos

Short version

slide-31
SLIDE 31

Region of Convergence, Taylor-Dirichlet series:

Assuming (A) and (B) then Λ = {λn, µn}∞

n=1 satisfies

lim

n→∞

log n λn = 0 lim

n→∞

µn λn = 0. Valiron (1929) : consider the series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

Cn = max{|cn,k| : k = 0, 1, 2, . . . , µn−1} lim sup

n→∞

log Cn λn = ξ ∈ R, P−ξ := {z : ℜz < −ξ}. Then g(z) is an analytic function in the left half-plane P−ξ converging uniformly on compact subsets. We call the line ℜz = −ξ the abscissa of convergence for g(z). Question : is it True that in every interval having length greater than 2πd on the line ℜz = −ξ, the series has at least One singularity?

  • E. Zikkos

Short version

slide-32
SLIDE 32

Positive Answers to the Singularity Question

  • E. Zikkos

Short version

slide-33
SLIDE 33

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

  • E. Zikkos

Short version

slide-34
SLIDE 34

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

◮ Zikkos (2005 Complex Variables):

  • E. Zikkos

Short version

slide-35
SLIDE 35

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

◮ Zikkos (2005 Complex Variables):

If Λ belongs to a certain class denoted by U(d, 0), with a restriction on the coefficients, the answer is YES.

  • E. Zikkos

Short version

slide-36
SLIDE 36

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

◮ Zikkos (2005 Complex Variables):

If Λ belongs to a certain class denoted by U(d, 0), with a restriction on the coefficients, the answer is YES.

◮ O. A. Krivosheeva (2012 St. Petersburg Math. J. ):

  • E. Zikkos

Short version

slide-37
SLIDE 37

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

◮ Zikkos (2005 Complex Variables):

If Λ belongs to a certain class denoted by U(d, 0), with a restriction on the coefficients, the answer is YES.

◮ O. A. Krivosheeva (2012 St. Petersburg Math. J. ):

If the Krivosheev characteristic SΛ is Equal to 0,

  • E. Zikkos

Short version

slide-38
SLIDE 38

Positive Answers to the Singularity Question

Suppose that Λ = {λn, µn} satisfies (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

◮ Zikkos (2005 Complex Variables):

If Λ belongs to a certain class denoted by U(d, 0), with a restriction on the coefficients, the answer is YES.

◮ O. A. Krivosheeva (2012 St. Petersburg Math. J. ):

If the Krivosheev characteristic SΛ is Equal to 0, then the answer is YES.

  • E. Zikkos

Short version

slide-39
SLIDE 39

Another Positive Answer

◮ Zikkos (2018):

  • E. Zikkos

Short version

slide-40
SLIDE 40

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0,

  • E. Zikkos

Short version

slide-41
SLIDE 41

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES.

  • E. Zikkos

Short version

slide-42
SLIDE 42

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) :

  • E. Zikkos

Short version

slide-43
SLIDE 43

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0).

  • E. Zikkos

Short version

slide-44
SLIDE 44

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers

  • E. Zikkos

Short version

slide-45
SLIDE 45

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers and let µn = pn+1 − pn.

  • E. Zikkos

Short version

slide-46
SLIDE 46

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers and let µn = pn+1 − pn. Then Λ = {pn, µn} belongs to the class U(1, 0).

  • E. Zikkos

Short version

slide-47
SLIDE 47

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers and let µn = pn+1 − pn. Then Λ = {pn, µn} belongs to the class U(1, 0).

Theorem

The Taylor-Dirichlet series g(z) =

  • n=1

µn−1

  • k=0

zk

  • epnz,

cn,k ∈ C defines an analytic function in the half-plane {z : ℜz < 0}

  • E. Zikkos

Short version

slide-48
SLIDE 48

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers and let µn = pn+1 − pn. Then Λ = {pn, µn} belongs to the class U(1, 0).

Theorem

The Taylor-Dirichlet series g(z) =

  • n=1

µn−1

  • k=0

zk

  • epnz,

cn,k ∈ C defines an analytic function in the half-plane {z : ℜz < 0} and it has at least One singularity

  • E. Zikkos

Short version

slide-49
SLIDE 49

Another Positive Answer

◮ Zikkos (2018):

If Λ belongs to the class U(d, 0), then the Krivosheev characteristic SΛ = 0, hence the answer is YES. Examples in U(d, 0) : (1) If (A) and (B) hold and µn = O(1), then Λ ∈ U(d, 0). (2) Let {pn} be the prime numbers and let µn = pn+1 − pn. Then Λ = {pn, µn} belongs to the class U(1, 0).

Theorem

The Taylor-Dirichlet series g(z) =

  • n=1

µn−1

  • k=0

zk

  • epnz,

cn,k ∈ C defines an analytic function in the half-plane {z : ℜz < 0} and it has at least One singularity in every open interval of length exceeding 2π and lying on the Imaginary axis.

  • E. Zikkos

Short version

slide-50
SLIDE 50

A Negative Answer

  • E. Zikkos

Short version

slide-51
SLIDE 51

A Negative Answer

Zikkos ( Ufa Math J.):

  • E. Zikkos

Short version

slide-52
SLIDE 52

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0,

  • E. Zikkos

Short version

slide-53
SLIDE 53

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0, there exists a multiplicity sequence Λ = {λn, µn} with µn → ∞, such that

  • E. Zikkos

Short version

slide-54
SLIDE 54

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0, there exists a multiplicity sequence Λ = {λn, µn} with µn → ∞, such that (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated).

  • E. Zikkos

Short version

slide-55
SLIDE 55

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0, there exists a multiplicity sequence Λ = {λn, µn} with µn → ∞, such that (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated). (C) SΛ < 0

  • E. Zikkos

Short version

slide-56
SLIDE 56

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0, there exists a multiplicity sequence Λ = {λn, µn} with µn → ∞, such that (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated). (C) SΛ < 0 and hence (Krivosheeva 2012 St. Petersburg Math. J.):

  • E. Zikkos

Short version

slide-57
SLIDE 57

A Negative Answer

Zikkos ( Ufa Math J.): for every d ≥ 0, there exists a multiplicity sequence Λ = {λn, µn} with µn → ∞, such that (A) Λ has Density d : lim

t→∞

  • λn≤t µn

t = d < ∞, (B) λn+1 − λn > c > 0, (Uniformly Separated). (C) SΛ < 0 and hence (Krivosheeva 2012 St. Petersburg Math. J.): there exists a Taylor-Dirichlet series such that it Can be Continued Analytically across the abscissa of convergence.

  • E. Zikkos

Short version

slide-58
SLIDE 58

The class U(d, 0)

  • E. Zikkos

Short version

slide-59
SLIDE 59

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) :

  • E. Zikkos

Short version

slide-60
SLIDE 60

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0.

  • E. Zikkos

Short version

slide-61
SLIDE 61

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c.

  • E. Zikkos

Short version

slide-62
SLIDE 62

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

  • E. Zikkos

Short version

slide-63
SLIDE 63

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

Choose a point in B(an, |an|α) ∩ R, call it bn, in an almost arbitrary way,

  • E. Zikkos

Short version

slide-64
SLIDE 64

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

Choose a point in B(an, |an|α) ∩ R, call it bn, in an almost arbitrary way, such that for all n = m either (I) bm = bn

  • E. Zikkos

Short version

slide-65
SLIDE 65

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

Choose a point in B(an, |an|α) ∩ R, call it bn, in an almost arbitrary way, such that for all n = m either (I) bm = bn

  • r

(II) |bm − bn| ≥ δ.

  • E. Zikkos

Short version

slide-66
SLIDE 66

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

Choose a point in B(an, |an|α) ∩ R, call it bn, in an almost arbitrary way, such that for all n = m either (I) bm = bn

  • r

(II) |bm − bn| ≥ δ. Rename {bn} into Λ = {λn, µn}.

  • E. Zikkos

Short version

slide-67
SLIDE 67

The class U(d, 0)

Zikkos (2005 Complex Variables, 2010 CMFT) : Consider a strictly increasing sequence {an} of positive real numbers, having density d with uniformly separated terms n/an → d, an+1 − an > c > 0. Choose two positive numbers α < 1, δ < c. For each term an consider the closed disk B(an, |an|α) = {z : |z − an| ≤ aα

n }.

Choose a point in B(an, |an|α) ∩ R, call it bn, in an almost arbitrary way, such that for all n = m either (I) bm = bn

  • r

(II) |bm − bn| ≥ δ. Rename {bn} into Λ = {λn, µn}. Then we say that Λ ∈ U(d, 0).

  • E. Zikkos

Short version

slide-68
SLIDE 68

The Class U(d, 0)

  • E. Zikkos

Short version

slide-69
SLIDE 69

The Class U(d, 0)

an

slide-70
SLIDE 70

The Class U(d, 0)

an R = aα

n

slide-71
SLIDE 71

The Class U(d, 0)

an R = aα

n

bn

slide-72
SLIDE 72

The Class U(d, 0)

an R = aα

n

bn an+1

slide-73
SLIDE 73

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

slide-74
SLIDE 74

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1

slide-75
SLIDE 75

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2

slide-76
SLIDE 76

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2 R = aα

n+2

slide-77
SLIDE 77

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2 R = aα

n+2

= bn+2

slide-78
SLIDE 78

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2 R = aα

n+2

= bn+2 an+3

slide-79
SLIDE 79

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2 R = aα

n+2

= bn+2 an+3 R = aα

n+3

slide-80
SLIDE 80

The Class U(d, 0)

an R = aα

n

bn an+1 R = aα

n+1

bn+1 an+2 R = aα

n+2

= bn+2 an+3 R = aα

n+3

= bn+3

  • E. Zikkos

Short version

slide-81
SLIDE 81

Singularities of Taylor-Dirichlet series

  • E. Zikkos

Short version

slide-82
SLIDE 82

Singularities of Taylor-Dirichlet series

Theorem A

Let the multiplicity-sequence Λ = {λn, µn}∞

n=1 belong to the class U(d, 0)

for some d > 0, and consider the Taylor-Dirichlet series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

cn,k ∈ C lim sup

n→∞

log Cn λn = ξ ∈ R, where Cn = max{|cn,k| : k = 0, 1, . . . , µn−1}.

  • E. Zikkos

Short version

slide-83
SLIDE 83

Singularities of Taylor-Dirichlet series

Theorem A

Let the multiplicity-sequence Λ = {λn, µn}∞

n=1 belong to the class U(d, 0)

for some d > 0, and consider the Taylor-Dirichlet series g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

cn,k ∈ C lim sup

n→∞

log Cn λn = ξ ∈ R, where Cn = max{|cn,k| : k = 0, 1, . . . , µn−1}. Then g(z) defines an analytic function in the half-plane {z : ℜz < −ξ} and it has at least One singularity in every open interval of length exceeding 2πd and lying on the line ℜz = −ξ.

  • E. Zikkos

Short version

slide-84
SLIDE 84

Second Goal

  • E. Zikkos

Short version

slide-85
SLIDE 85

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

  • E. Zikkos

Short version

slide-86
SLIDE 86

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

Characterize the closed span of the exponential system EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}

  • E. Zikkos

Short version

slide-87
SLIDE 87

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

Characterize the closed span of the exponential system EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1} in Lp(l) spaces where l is a simple closed rectifiable curve in C, and Gl is the domain bounded by the curve.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

  • E. Zikkos

Short version

slide-88
SLIDE 88

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

Characterize the closed span of the exponential system EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1} in Lp(l) spaces where l is a simple closed rectifiable curve in C, and Gl is the domain bounded by the curve.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

If f is in the closed span of EΛ in Lp(l),

  • E. Zikkos

Short version

slide-89
SLIDE 89

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

Characterize the closed span of the exponential system EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1} in Lp(l) spaces where l is a simple closed rectifiable curve in C, and Gl is the domain bounded by the curve.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

If f is in the closed span of EΛ in Lp(l), then f is in the Lp closure of polynomials,

  • E. Zikkos

Short version

slide-90
SLIDE 90

Second Goal

Given Λ = {λn, µn}∞

n=1 in U(d, 0)

Characterize the closed span of the exponential system EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1} in Lp(l) spaces where l is a simple closed rectifiable curve in C, and Gl is the domain bounded by the curve.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

If f is in the closed span of EΛ in Lp(l), then f is in the Lp closure of polynomials, hence f ∈ E p(Gl).

  • E. Zikkos

Short version

slide-91
SLIDE 91

Curve l is surrounded by a rectangle whose height is less than 2πd

  • E. Zikkos

Short version

slide-92
SLIDE 92

Curve l is surrounded by a rectangle whose height is less than 2πd

Height < 2πd

Theorem B

Suppose the Domain Gl bounded by the curve l is a Smirnov domain.

  • E. Zikkos

Short version

slide-93
SLIDE 93

Curve l is surrounded by a rectangle whose height is less than 2πd

Height < 2πd

Theorem B

Suppose the Domain Gl bounded by the curve l is a Smirnov domain. Suppose also that Λ = {λn, µn} has Density d.

  • E. Zikkos

Short version

slide-94
SLIDE 94

Curve l is surrounded by a rectangle whose height is less than 2πd

Height < 2πd

Theorem B

Suppose the Domain Gl bounded by the curve l is a Smirnov domain. Suppose also that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1

  • E. Zikkos

Short version

slide-95
SLIDE 95

Curve l is surrounded by a rectangle whose height is less than 2πd

Height < 2πd

Theorem B

Suppose the Domain Gl bounded by the curve l is a Smirnov domain. Suppose also that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 Coincides with the Smirnov space E p(Gl).

  • E. Zikkos

Short version

slide-96
SLIDE 96

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l).

  • E. Zikkos

Short version

slide-97
SLIDE 97

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l).

  • E. Zikkos

Short version

slide-98
SLIDE 98

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets.

  • E. Zikkos

Short version

slide-99
SLIDE 99

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets. ( B. Ya. Levin , A. F. Leont’ev):

  • E. Zikkos

Short version

slide-100
SLIDE 100

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets. ( B. Ya. Levin , A. F. Leont’ev): Since the density of Λ is d,

  • E. Zikkos

Short version

slide-101
SLIDE 101

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets. ( B. Ya. Levin , A. F. Leont’ev): Since the density of Λ is d, AND the height of the rectangle is less than 2πd,

  • E. Zikkos

Short version

slide-102
SLIDE 102

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets. ( B. Ya. Levin , A. F. Leont’ev): Since the density of Λ is d, AND the height of the rectangle is less than 2πd, then the system EΛ is Complete in H(K).

  • E. Zikkos

Short version

slide-103
SLIDE 103

Proof

It is enough to show that E p(Gl) is a subspace of the closed span of the exponential system EΛ in Lp(l). Since Gl is a Smirnov domain we have to show that the Lp closure of polynomials is a subspace of the closed span of the exponential system EΛ in Lp(l). Let H(K) be the space of functions analytic in the rectangle K with the topology of uniform convergence on compact subsets. ( B. Ya. Levin , A. F. Leont’ev): Since the density of Λ is d, AND the height of the rectangle is less than 2πd, then the system EΛ is Complete in H(K). Hence polynomials are approximated uniformly on the curve l by exponential polynomials.

  • E. Zikkos

Short version

slide-104
SLIDE 104

The curve l is Surrounding a rectangle whose height is 2πd

  • E. Zikkos

Short version

slide-105
SLIDE 105

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

  • E. Zikkos

Short version

slide-106
SLIDE 106

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d.

  • E. Zikkos

Short version

slide-107
SLIDE 107

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 is a Proper subspace

  • f the Smirnov space E p(Gl).
  • E. Zikkos

Short version

slide-108
SLIDE 108

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 is a Proper subspace

  • f the Smirnov space E p(Gl).

For any λ / ∈ {λn}, the function eλz does not belong to the closed span of the system.

  • E. Zikkos

Short version

slide-109
SLIDE 109

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 is a Proper subspace

  • f the Smirnov space E p(Gl).

For any λ / ∈ {λn}, the function eλz does not belong to the closed span of the system. Question:

  • E. Zikkos

Short version

slide-110
SLIDE 110

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 is a Proper subspace

  • f the Smirnov space E p(Gl).

For any λ / ∈ {λn}, the function eλz does not belong to the closed span of the system. Question: Can we characterize the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1?

  • E. Zikkos

Short version

slide-111
SLIDE 111

The curve l is Surrounding a rectangle whose height is 2πd

Height of rectangle ≥ 2πd

Theorem C

Suppose that Λ = {λn, µn} has Density d. Then the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1 is a Proper subspace

  • f the Smirnov space E p(Gl).

For any λ / ∈ {λn}, the function eλz does not belong to the closed span of the system. Question: Can we characterize the closed span of the exponential system EΛ in the space Lp(l) for p ≥ 1? We give an answer when Λ ∈ U(d, 0).

  • E. Zikkos

Short version

slide-112
SLIDE 112

Characterizing the closed span of EΛ

  • E. Zikkos

Short version

slide-113
SLIDE 113

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}.

  • E. Zikkos

Short version

slide-114
SLIDE 114

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd

slide-115
SLIDE 115

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd Sld the set of all such line segments

slide-116
SLIDE 116

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd Sld the set of all such line segments qld := sup{ℜz : ∀ z ∈ Sld}

slide-117
SLIDE 117

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd Sld the set of all such line segments qld := sup{ℜz : ∀ z ∈ Sld}

slide-118
SLIDE 118

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd Sld the set of all such line segments qld := sup{ℜz : ∀ z ∈ Sld} If f ∈ span(EΛ) in Lp(ld),

slide-119
SLIDE 119

Characterizing the closed span of EΛ

Let Λ belong to the class U(d, 0). Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Curve ld, Domain Gld length > 2πd Sld the set of all such line segments qld := sup{ℜz : ∀ z ∈ Sld} If f ∈ span(EΛ) in Lp(ld), f extends analytically in ℜz < qld as a Taylor-Dirichlet series

  • E. Zikkos

Short version

slide-120
SLIDE 120

The closed span of EΛ in Lp(ld)

  • E. Zikkos

Short version

slide-121
SLIDE 121

The closed span of EΛ in Lp(ld)

Theorem D

Let Λ = {λn, µn}∞

n=1 ∈ U(d, 0) and consider an ld curve and its qld

constant.

◮ Then every function f belonging to the closed span of EΛ in Lp(ld)

for p ≥ 1, not only extends analytically in the domain Gld and belongs to the Smirnov space E p(Gld).

◮ But it is also extended analytically in the half-plane

Hqld := {z : ℜz < qld}, admitting a unique Taylor-Dirichlet series representation of the form g(z) =

  • n=1

µn−1

  • k=0

cn,kzk

  • eλnz,

cn,k ∈ C, ∀ z ∈ Hqld with the series converging uniformly on compact subsets of Hqld .

  • E. Zikkos

Short version

slide-122
SLIDE 122

Crucial Tool: Distances in Lp(ld)

  • E. Zikkos

Short version

slide-123
SLIDE 123

Crucial Tool: Distances in Lp(ld)

Suppose that Λ = {λn, µn}∞

n=1 belongs to the class U(d, 0) and consider

an ld curve and its qld constant. Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}.

  • E. Zikkos

Short version

slide-124
SLIDE 124

Crucial Tool: Distances in Lp(ld)

Suppose that Λ = {λn, µn}∞

n=1 belongs to the class U(d, 0) and consider

an ld curve and its qld constant. Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Let pn,k(z) := zkeλnz And EΛn,k := EΛ \ {pn,k}.

  • E. Zikkos

Short version

slide-125
SLIDE 125

Crucial Tool: Distances in Lp(ld)

Suppose that Λ = {λn, µn}∞

n=1 belongs to the class U(d, 0) and consider

an ld curve and its qld constant. Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Let pn,k(z) := zkeλnz And EΛn,k := EΛ \ {pn,k}. Define the Distance between pn,k and the closed span of EΛn,k in Lp(ld) Dp,n,k := inf

g∈span(EΛn,k ) ||pn,k − g||Lp(ld)

  • E. Zikkos

Short version

slide-126
SLIDE 126

Crucial Tool: Distances in Lp(ld)

Suppose that Λ = {λn, µn}∞

n=1 belongs to the class U(d, 0) and consider

an ld curve and its qld constant. Let EΛ = {zkeλnz : n ∈ N, k = 0, 1, . . . , µn − 1}. Let pn,k(z) := zkeλnz And EΛn,k := EΛ \ {pn,k}. Define the Distance between pn,k and the closed span of EΛn,k in Lp(ld) Dp,n,k := inf

g∈span(EΛn,k ) ||pn,k − g||Lp(ld)

Theorem E

For every ǫ > 0 there is a constant uǫ > 0, independent of p ≥ 1, n ∈ N and k = 0, 1, . . . , µn − 1, but depending on Λ the curve ld, so that Dp,n,k ≥ uǫe(qld −ǫ)λn.

  • E. Zikkos

Short version

slide-127
SLIDE 127

A Biorthogonal sequence to EΛ in E 2(Gld) and a solution to a Moment Problem

Theorem F

◮ Let Λ = {λn, µn}∞ n=1 belong to the class U(d, 0) and consider an ld

curve and its qld constant.

  • E. Zikkos

Short version

slide-128
SLIDE 128

A Biorthogonal sequence to EΛ in E 2(Gld) and a solution to a Moment Problem

Theorem F

◮ Let Λ = {λn, µn}∞ n=1 belong to the class U(d, 0) and consider an ld

curve and its qld constant. Then there exists a family of functions {rn,k ∈ E 2(Gld) : n ∈ N, k = 0, 1, . . . , µn − 1} such that this family is the Unique Biorthogonal sequence to the system EΛ in E 2(Gld), belonging to span(EΛ) in E 2(Gld).

  • E. Zikkos

Short version

slide-129
SLIDE 129

A Biorthogonal sequence to EΛ in E 2(Gld) and a solution to a Moment Problem

Theorem F

◮ Let Λ = {λn, µn}∞ n=1 belong to the class U(d, 0) and consider an ld

curve and its qld constant. Then there exists a family of functions {rn,k ∈ E 2(Gld) : n ∈ N, k = 0, 1, . . . , µn − 1} such that this family is the Unique Biorthogonal sequence to the system EΛ in E 2(Gld), belonging to span(EΛ) in E 2(Gld).

◮ Moreover, for every ǫ > 0 there is a constant mǫ > 0, independent

  • f n and k, but depending on Λ and the curve ld, so that

||rn,k||E 2(Gld ) ≤ mǫe(−qld +ǫ)λn, ∀ n ∈ N, k = 0, 1, . . . , µn − 1.

  • E. Zikkos

Short version

slide-130
SLIDE 130

◮ Let {dn,k : n ∈ N, k = 0, 1, . . . , µn − 1} be a doubly-indexed

sequence of complex numbers such that lim sup

n→∞

log An λn < qld where An = max{|dn,k| : k = 0, 1, . . . , µn−1}.

  • E. Zikkos

Short version

slide-131
SLIDE 131

◮ Let {dn,k : n ∈ N, k = 0, 1, . . . , µn − 1} be a doubly-indexed

sequence of complex numbers such that lim sup

n→∞

log An λn < qld where An = max{|dn,k| : k = 0, 1, . . . , µn−1}. Then the function f (z) :=

  • n=1

µn−1

  • k=0

dn,krn,k(z)

  • E. Zikkos

Short version

slide-132
SLIDE 132

◮ Let {dn,k : n ∈ N, k = 0, 1, . . . , µn − 1} be a doubly-indexed

sequence of complex numbers such that lim sup

n→∞

log An λn < qld where An = max{|dn,k| : k = 0, 1, . . . , µn−1}. Then the function f (z) :=

  • n=1

µn−1

  • k=0

dn,krn,k(z)

  • belongs to E 2(Gld) and it is a solution to the moment problem
  • ld

zkeλnzf (z) |dz| = dn,k ∀ n ∈ N and k = 0, 1, 2, . . . µn − 1.

  • E. Zikkos

Short version

slide-133
SLIDE 133
  • A. S. Krivosheev, A fundamental principle for invariant subspaces in

convex domains, Izv. Ross. Acad. Nauk Ser. Mat. 68 no. 2 (2004), 71-136, English transl., Izv. Math. 68 no. 2 (2004), 291-353.

  • O. A. Krivosheeva, Singular points of the sum of a series of

exponential monomials on the boundary of the convergence domain, Algebra i Analiz 23 no. 2 (2011), 162-205; English transl., St. Petersburg Math. J. 23 no. 2 (2012), 321-350.

  • O. A. Krivosheeva, A. S. Krivosheev, Singular Points of the Sum of a

Dirichlet Series on the Convergence Line, Funktsional. Anal. i

  • Prilozhen. 49, no. 2 (2015), 54-69; English transl., Funct. Anal.
  • Appl. 49 no. 2 (2015), 122-134.

G, Polya, On converse gap theorems. Trans. Amer. Math. Soc. 52, (1942). 65-71.

  • M. G. Valiron, Sur les solutions des ´

equations diff´ erentielles lin´ eaires d’ordre infini et a coefficients constants, Ann. Ecole Norm. (3) 46 (1929), 25-53.

  • E. Zikkos

Short version

slide-134
SLIDE 134
  • E. Zikkos, On a theorem of Norman Levinson and a variation of the

Fabry Gap theorem, Complex Var. and Ell. Eqns. 50 no. 4 (2005), 229-255.

  • E. Zikkos, Analytic continuation of Taylor-Dirichlet series and

non-vanishing solutions of a differential equation of infinite order, CMFT 10 no. 1 (2010), 367-398.

  • E. Zikkos, A Taylor-Dirichlet series with no singularities on its

abscissa of convergence.

  • E. Zikkos

Short version

slide-135
SLIDE 135

THANK YOU VERY MUCH!!! ΣΑΣ ΕYΧΑΡΙΣΤΩ ΠΑΡ Α ΠΟΛY !!!

  • E. Zikkos

Short version