Jump to content

Phragmén–Lindelöf principle: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
m Example of application: fix wording and spaces
References: split authors in ref
 
(25 intermediate revisions by 11 users not shown)
Line 1: Line 1:
In [[complex analysis]], the '''Phragmén–Lindelöf principle''' (or '''method'''), first formulated by [[Lars Edvard Phragmén]] (1863–1937) and [[Ernst Leonard Lindelöf]] (1870-1946) in 1908, is a technique which employs an auxiliary, parameterized function to prove the boundedness of a holomorphic function <math>f</math> (i.e, <math>|f(z)|<M\ \ (z\in \Omega)</math>) on an unbounded domain <math>\Omega</math> when an additional (usually mild) condition constraining the growth of <math>|f|</math> on <math>\Omega</math> is given. It is a generalization of the [[maximum modulus principle]], which is only applicable to bounded domains.
In [[complex analysis]], the '''Phragmén–Lindelöf principle''' (or '''method'''), first formulated by [[Lars Edvard Phragmén]] (1863–1937) and [[Ernst Leonard Lindelöf]] (1870–1946) in 1908, is a technique which employs an auxiliary, parameterized function to prove the boundedness of a holomorphic function <math>f</math> (i.e, <math>|f(z)|<M\ \ (z\in \Omega)</math>) on an unbounded domain <math>\Omega</math> when an additional (usually mild) condition constraining the growth of <math>|f|</math> on <math>\Omega</math> is given. It is a generalization of the [[maximum modulus principle]], which is only applicable to bounded domains.


==Background==
==Background==


In the theory of complex functions, it is known that the [[Absolute value|modulus]] (absolute value) of a [[Holomorphic function|holomorphic]] (complex differentiable) function in the interior of a ''[[Bounded set|'''bounded''']]'' region is bounded by its modulus on the boundary of the region. More precisely, if a non-constant function <math>f:\mathbb{C}\to\mathbb{C}</math> is holomorphic in a bounded region<ref>The term ''region'' is not uniformly employed in the literature; here, a ''region'' is taken to mean a nonempty connected open subset of the complex plane.</ref> <math>\Omega</math> and [[continuous function|continuous]] on its closure <math>\overline{\Omega}=\Omega\cup\partial \Omega</math>, then <math display="inline">|f(z_0)|<\sup_{z\in \partial \Omega} |f(z)|</math> for all <math>z_0\in \Omega</math>. This is known as the ''maximum modulus principle.'' (In fact, since <math>\overline{\Omega}</math> is compact and <math>|f|</math> is continuous, there actually exists some <math>w_0\in\partial \Omega</math> such that <math display="inline">|f(w_0)|=\sup_{z\in \partial \Omega} |f(z)|</math>.) The maximum modulus principle is generally used to conclude that a holomorphic function is bounded in a region after showing that it is bounded on its boundary.
In the theory of complex functions, it is known that the [[Absolute value|modulus]] (absolute value) of a [[Holomorphic function|holomorphic]] (complex differentiable) function in the interior of a ''[[Bounded set|'''bounded''']]'' region is bounded by its modulus on the boundary of the region. More precisely, if a non-constant function <math>f:\mathbb{C}\to\mathbb{C}</math> is holomorphic in a bounded region<ref>The term ''region'' is not uniformly employed in the literature; here, a ''region'' is taken to mean a nonempty connected open subset of the complex plane.</ref> <math>\Omega</math> and [[continuous function|continuous]] on its closure <math>\overline{\Omega}=\Omega\cup\partial \Omega</math>, then <math display="inline">|f(z_0)|<\sup_{z\in \partial \Omega} |f(z)|</math> for all <math>z_0\in \Omega</math>. This is known as the ''maximum modulus principle.'' (In fact, since <math>\overline{\Omega}</math> is compact and <math>|f|</math> is continuous, there actually exists some <math>w_0\in\partial \Omega</math> such that <math display="inline">|f(w_0)|=\sup_{z\in \Omega} |f(z)|</math>.) The maximum modulus principle is generally used to conclude that a holomorphic function is bounded in a region after showing that it is bounded on its boundary.


However, the maximum modulus principle cannot be applied to an unbounded region of the complex plane. As a concrete example, let us examine the behavior of the holomorphic function <math>f(z) = \exp(\exp(z))</math> in the unbounded strip
However, the maximum modulus principle cannot be applied to an unbounded region of the complex plane. As a concrete example, let us examine the behavior of the holomorphic function <math>f(z) = \exp(\exp(z))</math> in the unbounded strip


:<math>S = \Big\{z:\Im(z)\in \big(-\frac{\pi}{2},\frac{\pi}{2}\big)\Big\}</math>.
:<math>S = \left\{z:\Im(z)\in \left(-\frac{\pi}{2},\frac{\pi}{2}\right)\right\}</math>.


Although <math>|f(x\pm \pi i/2)|=1</math>, so that <math>|f|</math> is bounded on boundary <math>\partial S</math>, <math>|f|</math> grows rapidly without bound when <math>|z|\to\infty</math> along the positive real axis. The difficulty here stems from the extremely fast growth of <math>|f|</math> along the positive real axis. If the growth rate of <math>|f|</math> is guaranteed to not be "too fast," as specified by an appropriate growth condition, the ''Phragmén–Lindelöf principle'' can be applied to show that boundedness of <math>f</math> on the region's boundary implies that <math>f</math> is in fact bounded in the whole region, effectively extending the maximum modulus principle to unbounded regions.
Although <math>|f(x\pm \pi i/2)|=1</math>, so that <math>|f|</math> is bounded on boundary <math>\partial S</math>, <math>|f|</math> grows rapidly without bound when <math>|z|\to\infty</math> along the positive real axis. The difficulty here stems from the extremely fast growth of <math>|f|</math> along the positive real axis. If the growth rate of <math>|f|</math> is guaranteed to not be "too fast," as specified by an appropriate growth condition, the ''Phragmén–Lindelöf principle'' can be applied to show that boundedness of <math>f</math> on the region's boundary implies that <math>f</math> is in fact bounded in the whole region, effectively extending the maximum modulus principle to unbounded regions.


== Outline of the technique ==
''The technique:'' Suppose we are given a holomorphic function <math>f</math> and an unbounded region <math>S</math>. In a typical Phragmén–Lindelöf argument, we introduce a certain multiplicative factor <math>h_\epsilon</math> satisfying <math display="inline">\lim_{\epsilon \to 0} h_\epsilon= 1</math> to "subdue" the growth of <math>f</math>, such that <math>|fh_\epsilon|<M</math> on the boundary of a bounded subregion <math>S_{\mathrm{bdd}}\subset S</math>. This allows us to apply the maximum modulus principle to <math>fh_\epsilon</math> and conclude that <math>fh_\epsilon</math> is bounded on <math>S_{\mathrm{bdd}}</math>. We then argue that the subregion could be expanded so as to encompass all points in <math>S</math>, establishing the boundedness of <math>fh_\epsilon</math> on <math>S</math>. Finally, we let <math>\epsilon\to 0</math> so that <math>fh_\epsilon\to f</math> in order to conclude that <math>f</math> must also be bounded on <math>S</math>.
Suppose we are given a holomorphic function <math>f</math> and an unbounded region <math>S</math>, and we want to show that <math>|f|\leq M</math> on <math>S</math>. In a typical Phragmén–Lindelöf argument, we introduce a certain multiplicative factor <math>h_\epsilon</math> satisfying <math display="inline">\lim_{\epsilon \to 0} h_\epsilon= 1</math> to "subdue" the growth of <math>f</math>. In particular, <math>h_\epsilon</math> is chosen such that (i): <math>fh_\epsilon</math> is holomorphic for all <math>\epsilon>0</math> and <math>|fh_\epsilon|\leq M</math> on the boundary <math>\partial S_{\mathrm{bdd}}</math> of an appropriate ''bounded'' subregion <math>S_{\mathrm{bdd}}\subset S</math>; and (ii): the asymptotic behavior of <math>fh_\epsilon</math> allows us to establish that <math>|fh_\epsilon|\leq M</math> for <math>z\in S\setminus \overline{S_{\mathrm{bdd}}}</math> (i.e., the unbounded part of <math>S</math> outside the closure of the bounded subregion). This allows us to apply the maximum modulus principle to first conclude that <math>|fh_\epsilon|\leq M</math> on <math>\overline{S_{\mathrm{bdd}}}</math> and then extend the conclusion to all <math>z\in S</math>. Finally, we let <math>\epsilon\to 0</math> so that <math>f(z)h_\epsilon(z)\to f(z)</math> for every <math>z\in S</math> in order to conclude that <math>|f|\leq M</math> on <math>S</math>.


In the literature of complex analysis, there are many examples of the Phragmén–Lindelöf principle applied to unbounded regions of differing types, and also a version of this principle may be applied in a similar fashion to [[subharmonic function|subharmonic]] and superharmonic functions.
In the literature of complex analysis, there are many examples of the Phragmén–Lindelöf principle applied to unbounded regions of differing types, and also a version of this principle may be applied in a similar fashion to [[subharmonic function|subharmonic]] and superharmonic functions.
Line 18: Line 19:
To continue the example above, we can impose a growth condition on a holomorphic function <math>f</math> that prevents it from "blowing up" and allows the Phragmén–Lindelöf principle to be applied. To this end, we now include the condition that
To continue the example above, we can impose a growth condition on a holomorphic function <math>f</math> that prevents it from "blowing up" and allows the Phragmén–Lindelöf principle to be applied. To this end, we now include the condition that


:<math>|f(z)|<\exp\big(A\exp(c\cdot|\Re(z)|)\big)</math>
:<math>|f(z)|<\exp\left(A\exp(c \cdot \left|\Re(z)\right|)\right)</math>


for some real constants <math>c<1</math> and <math>A<\infty</math>, for all <math>z\in S</math>. It can then be shown that <math>|f(z)|\leq 1</math> for all <math>z\in\partial S</math> implies that <math>|f(z)|\leq 1</math> in fact holds for all <math>z\in S</math>. Thus, we have the following proposition:
for some real constants <math>c<1</math> and <math>A<\infty</math>, for all <math>z\in S</math>. It can then be shown that <math>|f(z)|\leq 1</math> for all <math>z\in\partial S</math> implies that <math>|f(z)|\leq 1</math> in fact holds for all <math>z\in S</math>. Thus, we have the following proposition:
Line 24: Line 25:
'''Proposition.''' ''Let''
'''Proposition.''' ''Let''


:'''''<math>S=\Big\{z:\Im(z)\in \big(-\frac{\pi}{2},\frac{\pi}{2}\big)\Big\},\quad \overline{S}=\Big\{z:\Im(z)\in \big[-\frac{\pi}{2},\frac{\pi}{2}\big]\Big\}</math>'''''.
:<math>S=\left\{z:\Im(z)\in \left(-\frac{\pi}{2},\frac{\pi}{2}\right)\right\},\quad \overline{S}=\left\{z:\Im(z)\in \left[-\frac{\pi}{2},\frac{\pi}{2}\right]\right\}.</math>


''Let'' <math>f</math> ''be holomorphic on <math>S</math>'' ''and continuous on <math>\overline{S}</math>, and suppose there exist real constants <math>c<1,\ A<\infty</math>'' ''such that''
''Let'' <math>f</math> ''be holomorphic on <math>S</math> and continuous on <math>\overline{S}</math>, and suppose there exist real constants <math>c<1,\ A<\infty</math> such that''
:'''''<math>|f(z)|<\exp\big(A\exp(c\cdot|\Re(z)|)\big)</math>'''''
:<math>|f(z)|<\exp\bigl(A\exp(c\cdot|\Re(z)|)\bigr)</math>


''for all <math>z\in S</math>'' ''and <math>|f(z)|\leq 1</math> for all <math>z\in\overline{S}\setminus S=\partial S</math>. Then'' ''<math>|f(z)|\leq 1</math>'' ''for all <math>z\in S</math>''.
''for all <math>z\in S</math> and <math>|f(z)|\leq 1</math> for all <math>z\in\overline{S}\setminus S=\partial S</math>. Then <math>|f(z)|\leq 1</math> for all <math>z\in S</math>''.


Note that this conclusion fails when <math>c=1</math>, precisely as the motivating counterexample in the previous section demonstrates. The proof of this statement employs a typical Phragmén–Lindelöf argument:<ref>{{Cite book|url=https://archive.org/details/RudinW.RealAndComplexAnalysis3e1987|title=Real and Complex Analysis|last=Rudin|first=Walter|publisher=McGraw-Hill|year=1987|isbn=0070542341|location=New York|pages=257–259|quote=|via=}}</ref>
Note that this conclusion fails when <math>c=1</math>, precisely as the motivating counterexample in the previous section demonstrates. The proof of this statement employs a typical Phragmén–Lindelöf argument:<ref>{{Cite book|url=https://archive.org/details/RudinW.RealAndComplexAnalysis3e1987|title=Real and Complex Analysis|last=Rudin|first=Walter|publisher=McGraw-Hill |year=1987|isbn=0070542341|location=New York|pages=257–259}}</ref>


'''Proof:''' ''(Sketch)'' We fix <math>b\in(c,1)</math> and define for each <math>\epsilon>0</math> the auxiliary function <math>h_\epsilon</math> by <math display="inline">h_\epsilon(z)=e^{-\epsilon(e^{b z}+e^{-b z})}</math>. Moreover, for given <math>a>0</math>, we define <math>S_{a}</math> to be the open rectangle in the complex plane bounded by the vertices <math>\{a\pm i\pi/2,-a\pm i\pi/2\}</math>. Now, consider the function <math>fh_\epsilon</math>. It can be shown that <math>|f(z)h_\epsilon(z)|\to 0</math> as <math>|\Re(z)|\to\infty</math>. This allow us to find an <math>x_0</math> such that <math>|f(z)h_\epsilon(z)|\leq1</math> whenever <math>z\in\overline{S}</math> and <math>|\Re(z)|\geq x_0</math>. Because <math>S_{x_0}</math> is a bounded region, and <math>|f(z)h_\epsilon(z)|\leq 1</math> for all <math>z\in\partial S_{x_0}</math>, the maximum modulus principle implies that <math>|f(z)h_\epsilon(z)|\leq 1</math> for all <math>z\in S_{x_0}</math>. In fact, given any <math>x>x_0</math>, <math>|f(z)h_\epsilon(z)|\leq1</math> must still hold whenever <math>z\in\overline{S}</math> and <math>|\Re(z)|\geq x</math>. Thus, the same maximum modulus principle argument now implies that for any given <math>x>x_0</math> and <math>z\in S_{x}</math>, we have <math>|f(z)h_\epsilon(z)|\leq1</math>. Therefore, <math>|f(z)h_\epsilon(z)|\leq1</math> for all <math>z\in S</math>. Finally, because <math>fh_\epsilon\to f</math> as <math>\epsilon\to 0</math>, we conclude that <math>|f(z)|\leq 1</math> for all <math>z\in S</math>. [[Tombstone (typography)|<big>∎</big>]]
'''Proof:''' ''(Sketch)'' We fix <math>b\in(c,1)</math> and define for each <math>\epsilon>0</math> the auxiliary function <math>h_\epsilon</math> by <math display="inline">h_\epsilon(z)=e^{-\epsilon(e^{b z}+e^{-b z})}</math>. Moreover, for a given <math>a>0</math>, we define <math>S_{a}</math> to be the open rectangle in the complex plane enclosed within the vertices <math>\{a\pm i\pi/2,-a\pm i\pi/2\}</math>. Now, fix <math>\epsilon>0</math> and consider the function <math>fh_\epsilon</math>. Because one can show that <math>|h_\epsilon(z)|\leq1</math> for all <math>z\in \overline{S}</math>, it follows that <math>|f(z)h_\epsilon(z)|\leq 1</math> for <math>z\in\partial S</math>. Moreover, one can show for <math>z\in\overline{S}</math> that <math>|f(z)h_\epsilon(z)|\to 0</math> uniformly as <math>|\Re(z)|\to\infty</math>. This allows us to find an <math>x_0</math> such that <math>|f(z)h_\epsilon(z)|\leq1</math> whenever <math>z\in\overline{S}</math> and <math>|\Re(z)|\geq x_0</math>. Now consider the bounded rectangular region <math>S_{x_0}</math>. We have established that <math>|f(z)h_\epsilon(z)|\leq 1</math> for all <math>z\in\partial S_{x_0}</math>. Hence, the maximum modulus principle implies that <math>|f(z)h_\epsilon(z)|\leq 1</math> for all <math>z\in \overline{S_{x_0}}</math>. Since <math>|f(z)h_\epsilon(z)|\leq1</math> also holds whenever <math>z\in S</math> and <math>|\Re(z)|> x_0</math>, we have in fact shown that <math>|f(z)h_\epsilon(z)|\leq1</math> holds for all <math>z\in S</math>. Finally, because <math>fh_\epsilon\to f</math> as <math>\epsilon\to 0</math>, we conclude that <math>|f(z)|\leq 1</math> for all <math>z\in S</math>. [[Q.E.D.]]


== Phragmén–Lindelöf principle for a sector in the complex plane ==
== Phragmén–Lindelöf principle for a sector in the complex plane ==
A particularly useful statement proved using the Phragmén–Lindelöf principle bounds holomorphic functions on a sector of the complex plane if it is bounded on its boundary. This statement can be used to give a complex analytic proof of the [[G. H. Hardy|Hardy's]] [[Fourier transform#Uncertainty principle|uncertainty principle]], which states that a function and its Fourier transform cannot both decay faster than exponentially.<ref>{{Cite web|url=https://terrytao.wordpress.com/2009/02/18/hardys-uncertainty-principle/#more-1716|title=Hardy's Uncertainty Principle|last=Tao|first=Terence|authorlink=Terence Tao|date=2009-02-18|website=Updates on my research and expository papers, discussion of open problems, and other maths-related topics. By Terence Tao|access-date=}}</ref>
A particularly useful statement proved using the Phragmén–Lindelöf principle bounds holomorphic functions on a sector of the complex plane if it is bounded on its boundary. This statement can be used to give a complex analytic proof of the [[G. H. Hardy|Hardy's]] [[Fourier transform#Uncertainty principle|uncertainty principle]], which states that a function and its Fourier transform cannot both decay faster than exponentially.<ref>{{Cite web|url=https://terrytao.wordpress.com/2009/02/18/hardys-uncertainty-principle/#more-1716|title=Hardy's Uncertainty Principle|last=Tao|first=Terence|author-link=Terence Tao| date=2009-02-18|website=Updates on my research and expository papers, discussion of open problems, and other maths-related topics. By Terence Tao}}</ref>


'''Proposition.''' ''Let <math>F</math> be a function that is [[Holomorphic function|holomorphic]] in a [[Circular sector|sector]]''
'''Proposition.''' ''Let <math>F</math> be a function that is [[Holomorphic function|holomorphic]] in a [[Circular sector|sector]]''


:''<math> S = \left\{ z \, \big| \, \alpha < \arg z < \beta \right\} </math>''
:<math> S = \left\{ z \, \big| \, \alpha < \arg z < \beta \right\} </math>


''of central angle <math>\beta-\alpha=\pi/\lambda</math>, and continuous on its boundary. If''
''of central angle <math>\beta-\alpha=\pi/\lambda</math>, and continuous on its boundary. If''


:{{NumBlk|:|<math>|F(z)| \leq 1</math>|{{EquationRef|1}}}}
{{NumBlk|:|<math>|F(z)| \leq 1</math>|{{EquationRef|1}}}}


''for <math>z\in\partial S</math>, and''
''for <math>z\in\partial S</math>, and''


:{{NumBlk|:|<math>|F(z)| \leq e^{C |z|^\rho}</math>|{{EquationRef|2}}}}
{{NumBlk|:|<math>|F(z)| \leq e^{C |z|^\rho}</math>|{{EquationRef|2}}}}

''for all <math>z\in S</math>, where <math>\rho\in[0,\lambda)</math> and <math>C>0</math>, then ({{EquationNote|1}}) holds also for all <math>z\in S</math>.''
''for all <math>z\in S</math>, where <math>\rho\in[0,\lambda)</math> and <math>C>0</math>, then <math>|F(z)| \leq 1</math> holds also for all <math>z\in S</math>.''


===Remarks===
===Remarks===
* The condition ({{EquationNote|2}}) can be relaxed to
The condition ({{EquationNote|2}}) can be relaxed to
{{NumBlk|:|<math>\liminf_{r \to \infty} \sup_{\alpha < \theta < \beta} \frac{\log|F(re^{i\theta})|}{r^\rho} = 0 \quad \text{for some} \quad 0 \leq \rho < \lambda~,</math>|{{EquationRef|3}}}}

:{{NumBlk|:|<math>\liminf_{r \to \infty} \sup_{\alpha < \theta < \beta} \frac{\log|F(re^{i\theta})|}{r^\rho} = 0 \quad \text{for some} \quad 0 \leq \rho < \lambda~,</math>|{{EquationRef|3}}}}

with the same conclusion.
with the same conclusion.


Line 63: Line 62:


[[Carlson's theorem]] is an application of the principle to functions bounded on the imaginary axis.
[[Carlson's theorem]] is an application of the principle to functions bounded on the imaginary axis.

== See also ==

* [[Hadamard three-lines theorem]]


==References==
==References==
Line 68: Line 71:


* {{cite journal
* {{cite journal
|author1=Phragmén, Lars Edvard |author2=Lindelöf, Ernst
| author1 = Phragmén, Lars Edvard | author2 = Lindelöf, Ernst
| title = Sur une extension d'un principe classique de l'analyse et sur quelques propriétés des fonctions monogènes dans le voisinage d'un point singulier
| title = Sur une extension d'un principe classique de l'analyse et sur quelques propriétés des fonctions monogènes dans le voisinage d'un point singulier
| journal = Acta Math.
| journal = Acta Math.
Line 77: Line 80:
| issn = 0001-5962
| issn = 0001-5962
| doi = 10.1007/BF02415450
| doi = 10.1007/BF02415450
|url = https://zenodo.org/record/2177451
|doi-access = free
}}
}}
* {{cite journal
* {{cite journal
| last = Riesz
| last = Riesz
| first = Marcel
| first = Marcel
|authorlink= Marcel Riesz
|author-link= Marcel Riesz
| title = Sur le principe de Phragmén-Lindelöf
| title = Sur le principe de Phragmén-Lindelöf
| journal = Proceedings of the Cambridge Philosophical Society
| journal = Proceedings of the Cambridge Philosophical Society
| volume = 20
| volume = 20
| year = 1920
| year = 1920
}} (Correction, vol. 21, 1921).
}} (Corr. {{cite journal | title = Sur le principe de Phragmén-Lindelöf | volume = 21 | year = 1921 }})
* {{cite book
* {{cite book
| last = Titchmarsh
| last = Titchmarsh
| first = Edward Charles
| first = Edward Charles
|authorlink= Edward Charles Titchmarsh
|author-link= Edward Charles Titchmarsh
| title = The Theory of Functions
| title = The Theory of Functions
| edition = Second
| edition = Second
Line 99: Line 104:
* {{springer|author=E.D. Solomentsev|title=Phragmén–Lindelöf theorem|id=p/p072630}}
* {{springer|author=E.D. Solomentsev|title=Phragmén–Lindelöf theorem|id=p/p072630}}
* {{cite book
* {{cite book
| author = [[Elias M. Stein|Stein, Elias M.]] and Shakarchi, Rami
| last1 = Stein | first1 = Elias M. | author1-link = Elias M. Stein
| last2 = Shakarchi | first2 = Rami
| title = Complex analysis
| isbn = 0-691-11385-8
| series = Princeton Lectures in Analysis, II
|publisher = Princeton University Press
| location = Princeton, NJ
| location = Princeton, NJ
| publisher = Princeton University Press
| series = Princeton Lectures in Analysis, II
| year = 2003
| title = Complex analysis
| isbn = 0-691-11385-8
| year = 2003}}
}}


{{DEFAULTSORT:Phragmen-Lindelof principle}}
{{DEFAULTSORT:Phragmen-Lindelof principle}}

Latest revision as of 06:51, 1 May 2024

In complex analysis, the Phragmén–Lindelöf principle (or method), first formulated by Lars Edvard Phragmén (1863–1937) and Ernst Leonard Lindelöf (1870–1946) in 1908, is a technique which employs an auxiliary, parameterized function to prove the boundedness of a holomorphic function (i.e, ) on an unbounded domain when an additional (usually mild) condition constraining the growth of on is given. It is a generalization of the maximum modulus principle, which is only applicable to bounded domains.

Background

[edit]

In the theory of complex functions, it is known that the modulus (absolute value) of a holomorphic (complex differentiable) function in the interior of a bounded region is bounded by its modulus on the boundary of the region. More precisely, if a non-constant function is holomorphic in a bounded region[1] and continuous on its closure , then for all . This is known as the maximum modulus principle. (In fact, since is compact and is continuous, there actually exists some such that .) The maximum modulus principle is generally used to conclude that a holomorphic function is bounded in a region after showing that it is bounded on its boundary.

However, the maximum modulus principle cannot be applied to an unbounded region of the complex plane. As a concrete example, let us examine the behavior of the holomorphic function in the unbounded strip

.

Although , so that is bounded on boundary , grows rapidly without bound when along the positive real axis. The difficulty here stems from the extremely fast growth of along the positive real axis. If the growth rate of is guaranteed to not be "too fast," as specified by an appropriate growth condition, the Phragmén–Lindelöf principle can be applied to show that boundedness of on the region's boundary implies that is in fact bounded in the whole region, effectively extending the maximum modulus principle to unbounded regions.

Outline of the technique

[edit]

Suppose we are given a holomorphic function and an unbounded region , and we want to show that on . In a typical Phragmén–Lindelöf argument, we introduce a certain multiplicative factor satisfying to "subdue" the growth of . In particular, is chosen such that (i): is holomorphic for all and on the boundary of an appropriate bounded subregion ; and (ii): the asymptotic behavior of allows us to establish that for (i.e., the unbounded part of outside the closure of the bounded subregion). This allows us to apply the maximum modulus principle to first conclude that on and then extend the conclusion to all . Finally, we let so that for every in order to conclude that on .

In the literature of complex analysis, there are many examples of the Phragmén–Lindelöf principle applied to unbounded regions of differing types, and also a version of this principle may be applied in a similar fashion to subharmonic and superharmonic functions.

Example of application

[edit]

To continue the example above, we can impose a growth condition on a holomorphic function that prevents it from "blowing up" and allows the Phragmén–Lindelöf principle to be applied. To this end, we now include the condition that

for some real constants and , for all . It can then be shown that for all implies that in fact holds for all . Thus, we have the following proposition:

Proposition. Let

Let be holomorphic on and continuous on , and suppose there exist real constants such that

for all and for all . Then for all .

Note that this conclusion fails when , precisely as the motivating counterexample in the previous section demonstrates. The proof of this statement employs a typical Phragmén–Lindelöf argument:[2]

Proof: (Sketch) We fix and define for each the auxiliary function by . Moreover, for a given , we define to be the open rectangle in the complex plane enclosed within the vertices . Now, fix and consider the function . Because one can show that for all , it follows that for . Moreover, one can show for that uniformly as . This allows us to find an such that whenever and . Now consider the bounded rectangular region . We have established that for all . Hence, the maximum modulus principle implies that for all . Since also holds whenever and , we have in fact shown that holds for all . Finally, because as , we conclude that for all . Q.E.D.

Phragmén–Lindelöf principle for a sector in the complex plane

[edit]

A particularly useful statement proved using the Phragmén–Lindelöf principle bounds holomorphic functions on a sector of the complex plane if it is bounded on its boundary. This statement can be used to give a complex analytic proof of the Hardy's uncertainty principle, which states that a function and its Fourier transform cannot both decay faster than exponentially.[3]

Proposition. Let be a function that is holomorphic in a sector

of central angle , and continuous on its boundary. If

for , and

for all , where and , then holds also for all .

Remarks

[edit]

The condition (2) can be relaxed to

with the same conclusion.

Special cases

[edit]

In practice the point 0 is often transformed into the point ∞ of the Riemann sphere. This gives a version of the principle that applies to strips, for example bounded by two lines of constant real part in the complex plane. This special case is sometimes known as Lindelöf's theorem.

Carlson's theorem is an application of the principle to functions bounded on the imaginary axis.

See also

[edit]

References

[edit]
  1. ^ The term region is not uniformly employed in the literature; here, a region is taken to mean a nonempty connected open subset of the complex plane.
  2. ^ Rudin, Walter (1987). Real and Complex Analysis. New York: McGraw-Hill. pp. 257–259. ISBN 0070542341.
  3. ^ Tao, Terence (2009-02-18). "Hardy's Uncertainty Principle". Updates on my research and expository papers, discussion of open problems, and other maths-related topics. By Terence Tao.