# Lengths, areas and Lipschitz-type spaces of planar harmonic mappings

Description
Lengths, areas and Lipschitz-type spaces of planar harmonic mappings
Categories
Published

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
LENGTHS, AREAS AND LIPSCHITZ-TYPE SPACES OF PLANARHARMONIC MAPPINGS SH. CHEN, S. PONNUSAMY  † , AND A. RASILA Abstract.  In this paper, we establish a three circles type theorem, involving theharmonic area function, for harmonic mappings. Also, we give bounds for lengthand area distortion for harmonic quasiconformal mappings. Finally, we will studycertain Lipschitz-type spaces on harmonic mappings. 1.  Introduction and main results Let  D  be a simply connected subdomain of the complex plane  C . A complex-valued function  f   deﬁned in  D  is called a  harmonic mapping   in  D  if and only if both the real and the imaginary parts of   f   are real harmonic in  D . It is known thatevery harmonic mapping  f   deﬁned in  D  admits a decomposition  f   =  h  +  g , where h  and  g  are analytic in  D . Since the Jacobian  J  f   of   f   is given by J  f   =  | f  z | 2 −| f  z | 2 :=  | h ′ | 2 −| g ′ | 2 ,f   is locally univalent and sense-preserving in  D  if and only if   | g ′ ( z  ) |  <  | h ′ ( z  ) |  in D ; or equivalently if   h ′ ( z  )   = 0 and the dilatation  ω  =  g ′ /h ′  has the property that | ω ( z  ) |  <  1 in  D  (see ). Let H ( D ) denote the class of all sense-preserving harmonicmappings in  D . We refer to [9, 11] for basic results in the theory of planar harmonic mappings.For  a  ∈ C , let  D ( a,r ) =  { z   :  | z   − a |  < r } . In particular, we use  D r  to denote thedisk  D (0 ,r ) and  D , the open unit disk  D 1 .The classical theorem of three circles [1, 28], also called  Hadamard’s three circles theorem  , states that if   f   is an analytic function in the annulus  B ( r 1 ,r 2 ) =  { z   : 0  <r 1  <  | z  |  =  r < r 2  <  ∞} , continuous on  B ( r 1 ,r 2 ), and  M  1 ,  M  2  and  M   are themaxima of   f   on the three circles corresponding to  r 1 ,  r 2  and  r , respectively, then M  log  r 2 r 1  ≤  M  log  r 2 r 1  M  log  rr 1 2  . Equivalently, we can reformulate this result into a simpler form. That is if   f   isanalytic on the annulus  B ( r 1 , 1) =  { z   : 0  < r 1  <  | z  |  <  1 } , continuous on theclosure, and | f  ( z  ) | ≤  m  =  r α 1 ,  | z  |  =  r 1  and  | f  ( z  ) | ≤  1 ,  | z  |  = 1 , File: Ch-P-R-AreaLength.tex, printed: 15-9-2013, 13.11 2000  Mathematics Subject Classiﬁcation.  Primary: 30H05, 30H30; Secondary: 30C20, 30C45. Key words and phrases.  Harmonic mapping, three circles theorem, area function. † Corresponding author. This author is on leave from the Department ofMathematics, Indian Institute of Technology Madras, Chennai-600 036, India  . 1  2 Sh. Chen, S. Ponnusamy and A. Rasila then Hadamard’s result states that, for  r 1  ≤  r  ≤  1 , | f  ( z  ) | ≤  m log r log r 1  =  r α ,  | z  |  =  r, where  α  is an integer.The srcinal three circles theorem was given by Hadamard without proof in 1896, and comprehensive discussion about the history of this result can be found in[20, pp. 323–325] and . It is a natural question, what results of this type can beproved for other classes of functions and, indeed, there are numerous generalizationsof the thee circles theorem in the literature, see e.g. [3, 21, 26, 30]. In this paper, our ﬁrst aim is to establish an area version of the three circles theorem (cf. areaversion of Schwarz’ lemma ).For a harmonic mapping  f   in  D  and  r  ∈  [0 , 1), the  harmonic area function   S  f  ( r )of   f  , counting multiplicity, is deﬁned by S  f  ( r ) =   D r J  f  ( z  ) dσ ( z  ) , where  dσ  denotes the normalized Lebesgue area measure on D (cf. ). In particular,let S  f  (1) = sup 0 <r< 1 S  f  ( r ) . Theorem 1.  Let   f   =  h  +  g  be harmonic in   D , where   h  and   g  are analytic. If  S  f  ( r 1 )  ≤  m <  1 ,  S  f  (1)  ≤  1  and for all   n  ∈ { 1 , 2 ,... } ,  | g ( n ) (0) | ≤ | h ( n ) (0) | , then for  r 1  ≤  r <  1 , (1.1)  S  f  ( r )  ≤  m log r log r 1 . The estimate of   (1.1)  is sharp and the extremal function is   f  ( z  ) =  αz   +  βz  , where  α  and   β   are constant with   | α | 2 −| β  | 2 = 1 . Corollary 1.1.  Let   f   be analytic in   D  satisfying   S  f  ( r 1 )  ≤  m  and   S  f  (1)  ≤  1 ,  where  0  < r 1  <  1 . Then for   r 1  ≤  r <  1 , (1.2)  S  f  ( r )  ≤  m log r log r 1 . The estimate of   (1.2)  is sharp and the extremal function is   f  ( z  ) =  λz  , where   | λ |  = 1 are constant. For  p  ∈  (0 , ∞ ], the  harmonic Hardy space   h  p consists of all harmonic functions  f  such that   f    p  <  ∞ , where  f    p  =  sup 0 <r< 1 M   p ( r,f  ) if   p  ∈  (0 , ∞ ) , sup z ∈ D | f  ( z  ) |  if   p  =  ∞ ,  and  M   p p ( r,f  ) = 12 π    2 π 0 | f  ( re iθ ) |  p dθ. If   f   ∈  h  p for some  p >  0, then the radial limits f  ( e iθ ) = lim r → 1 − f  ( re iθ )exist for almost every  θ  ∈  [0 , 2 π ) (cf. ).  Lengths, areas and Lipschitz-type spaces of planar harmonic mappings 3 We recall that a function  f   ∈ H ( D ) is said to be  K  -quasiregular  ,  K   ∈  [1 , ∞ ), if for  z   ∈  D , Λ f  ( z  )  ≤  Kλ f  ( z  ). In addition, if   f   is univalent in  D , then  f   is called a K  -quasiconformal   harmonic mapping  D .Let Ω be a domain of   C , with non-empty boundary. Let  d Ω ( z  ) be the Euclideandistance from  z   to the boundary  ∂  Ω of Ω. In particular, we always use  d ( z  ) todenote the Euclidean distance from  z   to the boundary of   D .  The area of a set G  ⊂  C  is denoted by  A ( G ). The area problem of analytic functions has attractedmuch attention (see [2, 29, 31, 32]). We investigate the area problem of harmonic mappings and obtain the following result. Theorem 2.  Let   Ω 1  and   Ω 2  be two proper and simply connected subdomains of   C containing the point of srcin. Then for a sense-preserving and   K  -quasiconformal harmonic mapping   f   deﬁned in   Ω 1  with   f  (0) = 0 , (1.3)  KA  f  (Ω 1 ) ∩ Ω 2  +  A ( f  − 1 (Ω 2 ))  ≥  min { d 2Ω 1 (0) ,d 2Ω 2 (0) } . Moreover, if   K   = 1 , then the estimate of   (1.3)  is sharp. We remark that Theorem 2 is a generalization of [29, Theorem].For a harmonic mapping  f   deﬁned on D , we use the following standard notations:Λ f  ( z  ) = max 0 ≤ θ ≤ 2 π | f  z ( z  ) +  e − 2 iθ f  z ( z  ) |  =  | f  z ( z  ) | + | f  z ( z  ) | and λ f  ( z  ) = min 0 ≤ θ ≤ 2 π | f  z ( z  ) +  e − 2 iθ f  z ( z  ) |  =  | f  z ( z  ) |−| f  z ( z  ) |  . Further, a planar harmonic mapping  f   deﬁned on  D  is called a  harmonic Bloch mapping   if  β  f   = sup z,w ∈ D , z  = w | f  ( z  ) − f  ( w ) | ρ ( z,w )  <  ∞ . Here  β  f   is called the  Lipschitz number   of   f  , and ρ ( z,w ) = 12 log  1 + | ( z   − w ) / (1 − zw ) | 1 −| ( z   − w ) / (1 − zw ) |  = arctanh  z   − w 1 − zw  denotes the hyperbolic distance between  z   and  w  in  D . It is known that β  f   = sup z ∈ D  (1 −| z  | 2 )Λ f  ( z  )  . Clearly, a harmonic Bloch mapping  f   is uniformly continuous as a map betweenmetric spaces, f   : ( D ,ρ )  →  ( C , |·| ) , and for all  z,w  ∈ D  we have the Lipschitz inequality | f  ( z  ) − f  ( w ) | ≤  β  f   ρ ( z,w ) . A well-known fact is that the set of all harmonic Bloch mappings, denoted by thesymbol  HB  , forms a complex Banach space with the norm  ·  given by  f   HB  =  | f  (0) | + sup z ∈ D { (1 −| z  | 2 )Λ f  ( z  ) } .  4 Sh. Chen, S. Ponnusamy and A. Rasila Specially, we use  B   to denote the set of all analytic functions deﬁned in  D  whichforms a complex Banach space with the norm  f   B  =  | f  (0) | + sup z ∈ D { (1 −| z  | 2 ) | f  ′ ( z  ) |} . The reader is referred to [10, Theorem 2] (or [5, 6]) for a detailed discussion. For  r  ∈  [0 , 1), the length of the curve  C  ( r ) =  w  =  f  ( re iθ ) :  θ  ∈  [0 , 2 π ]  , countingmultiplicity, is deﬁned by l f  ( r ) =    2 π 0 | df  ( re iθ ) |  =  r    2 π 0  f  z ( re iθ ) − e − 2 iθ f  z ( re iθ )  dθ, where  f   is a harmonic mapping deﬁned in D . In particular, let  l f  (1) = sup 0 <r< 1 l f  ( r ). Theorem 3.  Let   f  ( z  ) =  ∞ n =0 a n z  n +  ∞ n =1 b n z  n be a sense-preserving   K  -quasiconformal harmonic mapping. If   l f  (1)  <  ∞ , then for   n  ≥  1 , (1.4)  | a n | + | b n | ≤  Kl f  (1)2 nπ and  (1.5) Λ f  ( z  )  ≤  l f  (1) √  K  2 π (1 −| z  | ) . Moreover,  f   ∈ HB   and   β  f   ≤  l f  (1) √  K π  .  In particular, if   K   = 1 , the estimates of   (1.4) and   (1.5)  are sharp, and the extremal function is   f  ( z  ) =  z. A continuous increasing function  ω  : [0 , ∞ )  →  [0 , ∞ ) with  ω (0) = 0 is called a majorant   if   ω ( t ) /t  is non-increasing for  t >  0. Given a subset Ω of   C , a function f   : Ω  →  C  is said to belong to the  Lipschitz space   L ω (Ω) if there is a positiveconstant  C   such that(1.6)  | f  ( z  ) − f  ( w ) | ≤  Cω ( | z   − w | ) for all  z, w  ∈  Ω . For  δ  0  >  0, let(1.7)    δ 0 ω ( t ) t dt  ≤  C   · ω ( δ  ) ,  0  < δ < δ  0 , and(1.8)  δ     + ∞ δ ω ( t ) t 2  dt  ≤  C   · ω ( δ  ) ,  0  < δ < δ  0 , where  ω  is a majorant and  C   is a positive constant.A majorant  ω  is said to be  regular   if it satisﬁes the conditions (1.7) and (1.8) (see[12, 13, 22, 23, 24]).Let  G  be a proper subdomain of   C . We say that a function  f   belongs to the local Lipschitz space   loc L ω ( G ) if (1.6) holds, with a ﬁxed positive constant  C  ,whenever  z   ∈  G  and  | z   − w |  <  12 d G ( z  ) (cf. [14, 18]). Moreover,  G  is said to be a L ω -extension domain   if   L ω ( G ) = loc L ω ( G ) .  The geometric characterization of   L ω -extension domains was ﬁrst given by Gehring and Martio . Then Lappalainen generalized their characterization, and proved that  G  is a  L ω -extension domain  Lengths, areas and Lipschitz-type spaces of planar harmonic mappings 5 if and only if each pair of points  z,w  ∈  G  can be joined by a rectiﬁable curve  γ   ⊂  G satisfying(1.9)   γ  ω ( d G ( z  )) d G ( z  )  ds ( z  )  ≤  Cω ( | z   − w | )with some ﬁxed positive constant  C   =  C  ( G,ω ), where  ds  stands for the arc lengthmeasure on  γ  . Furthermore, Lappalainen [18, Theorem 4.12] proved that  L ω -extension domains exist only for majorants  ω  satisfying (1.7). Theorem A.  ([16, Theorem 3])  f   ∈ B   if and only if  sup z,w ∈ D ,z  = w   (1 −| z  | 2 )(1 −| w | 2 ) | f  ( z  ) − f  ( w ) || z   − w |  <  ∞ . The following result is a generalization of Theorem A. For the related studies of this topic for real functions, we refer to [25, 27]. Theorem 4.  Let   f   be a harmonic mapping in   D  and   ω  be a majorant. Then the  following are equivalent: (a)  There exists a constant   C  1  >  0  such that for all   z   ∈ D , Λ f  ( z  )  ≤  C  1 ω   1 d ( z  )  ;(b)  There exists a constant   C  2  >  0  such that for all   z,w  ∈ D  with   z    =  w , | f  ( z  ) − f  ( w ) || z   − w | ≤  C  2 ω   1   d ( z  ) d ( w )  ;(c)  There exists a constant   C  3  >  0  such that for all   r  ∈  (0 ,d ( z  )] , 1 | D ( z,r ) |   B n ( z,r ) | f  ( ζ  ) − f  ( z  ) | dA ( ζ  )  ≤  C  3 rω  1 r  , where   dA  denotes the Lebesgue area measure in   D . Note that if   ω ( t ) =  t  and  f   is analytic, then (a) ⇐⇒ (b) in Theorem 4 implies thatTheorem A.Krantz  proved the following Hardy-Littlewood-type theorem for harmonicfunctions with respect to the majorant  ω ( t ) =  ω α ( t ) =  t α (0  < α  ≤  1). Theorem B.  ([17, Theorem 15.8])  Let   u  be a real harmonic function in   D  and  0  < α  ≤  1 . Then   u  satisﬁes  |∇ u ( z  ) | ≤  C ω α  d ( z  )  d ( z  )  for all   z   ∈ D if and only if  | u ( z  ) − u ( w ) | ≤  Cω α ( | z   − w | )  for all   z,w  ∈ D . We generalize Theorem B to the following form.

May 21, 2019

#### UNIVERSIDADE FEDERAL DO PARÁ INSTITUTO DE FILOSOFIA E CIÊNCIAS HUMANAS PROGRAMA DE PÓS-GRADUAÇÃO EM HISTÓRIA SOCIAL DA AMAZÔNIA

May 21, 2019
Search
Similar documents

View more...
Tags

## Finite Element Method

Related Search 