Chapter 2.14

Conformal Mapping

[33/0%]
Show
LaTeX
Problem 2.14.1

Find necessary and sufficient conditions which the complex numbers a,b,ca, b, c, and dd have to satisfy so that the linear fractional transformation z→(az+b)/(cz+d)z \rightarrow (a z+b) /(c z+d) maps the upper half plane onto itself.

?
Problem 2.14.2

In Theorem 11.14 the hypotheses were, in simplified form, that Ω⊂Π+,L\Omega \subset \Pi^{+}, L is on the real axis, and Im(f(z))→0\mathrm{Im}\left(f(z)\right) \rightarrow 0 as z→Lz \rightarrow L. Use this theorem to establish analogous reflection theorems under the following hypotheses:

?
(a)

Ω⊂Π+,L\Omega \subset \Pi^{+}, L on real axis, ∣f(z)∣→1\left|f(z)\right| \rightarrow 1 as z→Lz \rightarrow L.

(b)

Ω⊂U,L⊂T,∣f(z)∣→1\Omega \subset U, L \subset T,\left|f(z)\right| \rightarrow 1 as z→Lz \rightarrow L.

(c)

Ω⊂U,L⊂T,Im(f(z))→0\Omega \subset U, L \subset T, \mathrm{Im}\left(f(z)\right) \rightarrow 0 as z→Lz \rightarrow L.

In case (b), if ff has a zero at α∈Ω\alpha \in \Omega, show that its extension has a pole at 1/αˉ1 / \bar{\alpha }. What are the analogues of this in cases (a) and (c)?

Problem 2.14.3

Suppose RR is a rational function such that ∣R(z)∣=1\left|R(z)\right| = 1 if ∣z∣=1\left|z\right| = 1. Prove that

R(z)=czm∏n=1kz−αn1−αˉnz R(z) = c z^{m} \prod _{n = 1}^{k} \frac{z-\alpha _{n}}{1-\bar{\alpha }_{n} z}

where cc is a constant, mm is an integer, and α1,…,αk\alpha_{1}, \ldots , \alpha_{k} are complex numbers such that αn≠0\alpha_{n} \neq 0 and ∣αn∣≠1\left|\alpha_{n}\right| \neq 1. Note that each of the above factors has absolute value 1 if ∣z∣=1\left|z\right| = 1.

?
Problem 2.14.4

Obtain an analogous description of those rational functions which are positive on TT. Hint: Such a function must have the same number of zeros as poles in UU. Consider products of factors of the form

(z−α)(1−αˉz)(z−β)(1−βˉz) \frac{(z-\alpha )(1-\bar{\alpha } z)}{(z-\beta )(1-\bar{\beta } z)}

where ∣α∣<1\left|\alpha \right| < 1 and ∣β∣<1\left|\beta \right| < 1.

?
Problem 2.14.5

Suppose ff is a trigonometric polynomial,

f(θ)=∑k=−nnakeikθ f(\theta ) = \sum _{k = -n}^{n} a_{k} e^{i k \theta }

and f(θ)>0f(\theta ) > 0 for all real θ\theta. Prove that there is a polynomial P(z)=c0+c1z+⋯+cnznP(z) = c_{0}+c_{1} z+\cdots +c_{n} z^{n} such that

f(θ)=∣P(eiθ)∣2(θ real ). f(\theta ) = \left|P\left(e^{i \theta }\right)\right|^{2} \quad (\theta \text{ real }).

Hint: Apply Exercise 4 to the rational function Σakzk\Sigma a_{k} z^{k}. Is the result still valid if we assume f(θ)≥0f(\theta ) \geq 0 instead of f(θ)>0f(\theta ) > 0?

?
Problem 2.14.6

Find the fixed points of the mappings φα\varphi_{\alpha } (Definition 12.3). Is there a straight line which φα\varphi_{\alpha } maps to itself?

?
Problem 2.14.7

Find all complex numbers α\alpha for which fαf_{\alpha } is one-to-one in UU, where

fα(z)=z1+αz2 f_{\alpha }(z) = \frac{z}{1+\alpha z^{2}}

Describe fα(U)f_{\alpha }(U) for all these cases.

?
Problem 2.14.8

Suppose f(z)=z+(1/z)f(z) = z+(1 / z). Describe the families of ellipses and hyperbolas onto which ff maps circles with center at 0 and rays through 0.

?
Problem 2.14.9
?
(a)

Suppose Ω={z:−1<Re(z)<1}\Omega = \left\{ z:-1 < \mathrm{Re}\left(z\right) < 1\right\}. Find an explicit formula for the one-to-one conformal mapping ff of Ω\Omega onto UU for which f(0)=0f(0) = 0 and f′(0)>0f^{\prime }(0) > 0. Compute f′(0)f^{\prime }(0).

(b)

Note that the inverse of the function constructed in (a)(a) has its real part bounded in UU, whereas its imaginary part is unbounded. Show that this implies the existence of a continuous real function uu on Uˉ\bar{U} which is harmonic in UU and whose harmonic conjugate vv is unbounded in UU.

(c)

Suppose g∈H(U),∣Re(g)∣<1g \in H(U),\left|\mathrm{Re}\left(g\right)\right| < 1 in UU, and g(0)=0g(0) = 0. Prove that

∣g(reiθ)∣≤2πlog⁡1+r1−r \left|g\left(r e^{i \theta }\right)\right| \leq \frac{2}{\pi } \log \frac{1+r}{1-r}

Hint: See Exercise 10.

(d)

Let Ω\Omega be the strip that occurs in Theorem 12.9. Fix a point α+iβ\alpha +i \beta in Ω\Omega. Let hh be a conformal one-to-one mapping of Ω\Omega onto Ω\Omega that carries α+iβ\alpha +i \beta to 0. Prove that

∣h′(α+iβ)∣=1/cos⁡β \left|h^{\prime }(\alpha +i \beta )\right| = 1 / \cos \beta
Remark.
?

vv is the function which makes u+ivu+i v holomorphic in UU; we can determine vv uniquely by the requirement v(0)=0v(0) = 0.

Problem 2.14.10

Suppose ff and gg are holomorphic mappings of UU into Ω,f\Omega , f is one-to-one, f(U)=Ωf(U) = \Omega, and f(0)=g(0)f(0) = g(0). Prove that

g(D(0;r))⊂f(D(0;r))(0<r<1) g(D(0 ; r)) \subset f(D(0 ; r)) \quad (0 < r < 1)
?
Problem 2.14.11

Let Ω\Omega be the upper half of the unit disc UU. Find the conformal mapping ff of Ω\Omega onto UU that carries {−1,0,1}\left\{ -1,0,1\right\} to {−1,−i,1}\left\{ -1,-i, 1\right\}. Find z∈Ωz \in \Omega such that f(z)=0f(z) = 0. Find f(i/2)f(i / 2). Hint: f=φ∘s∘ψf = \varphi \circ s \circ \psi, where φ\varphi and ψ\psi are linear fractional transformations and s(λ)=λ2s(\lambda ) = \lambda^{2}.

?
Problem 2.14.12

Suppose Ω\Omega is a convex region, f∈H(Ω)f \in H(\Omega ), and Re(f′(z))>0\mathrm{Re}\left(f^{\prime }(z)\right) > 0 for all z∈Ωz \in \Omega. Prove that ff is one-to-one in Ω\Omega. Is the result changed if the hypothesis is weakened to Re(f′(z))≥0\mathrm{Re}\left(f^{\prime }(z)\right) \geq 0? (Exclude the trivial case f=f = constant). Show by an example that " convex" cannot be replaced by " simply connected".

?
Problem 2.14.13

Suppose Ω\Omega is a region, fn∈H(Ω)f_{n} \in H(\Omega ) for n=1,2,3,…n = 1,2,3, \ldots, each fnf_{n} is one-to-one in Ω\Omega, and fn→ff_{n} \rightarrow f uniformly on compact subsets of Ω\Omega. Prove that ff is either constant or one-to-one in Ω\Omega. Show that both cases can occur.

?
Problem 2.14.14

Suppose Ω={x+iy:−1<y<1},f∈H(Ω),∣f∣<1\Omega = \left\{ x+i y:-1 < y < 1\right\} , f \in H(\Omega ),\left|f\right| < 1, and f(x)→0f(x) \rightarrow 0 as x→∞x \rightarrow \infty. Prove that

lim⁡x→∞f(x+iy)=0(−1<y<1) \lim _{x \rightarrow \infty } f(x+i y) = 0 \quad (-1 < y < 1)

and that the passage to the limit is uniform if yy is confined to an interval [−α,α][-\alpha , \alpha ], where α<1\alpha < 1. Hint: Consider the sequence {fn}\left\{ f_{n}\right\}, where fn(z)=z+nf_{n}(z) = z+n, in the square ∣x∣<1,∣y∣<1\left|x\right| < 1,\left|y\right| < 1.

What does this theorem tell about the behavior of a function g∈H∞g \in H^{\infty } near a boundary point of UU at which the radial limit of gg exists?

?
Problem 2.14.15

Let F\mathscr {F} be the class of all f∈H(U)f \in H(U) such that Re(f)>0\mathrm{Re}\left(f\right) > 0 and f(0)=1f(0) = 1. Show that F\mathscr {F} is a normal family. Can the condition " f(0)=1f(0) = 1 " be omitted? Can it be replaced by " ∣f(0)∣≤1\left|f(0)\right| \leq 1 "?

?
Problem 2.14.16

Let F\mathscr {F} be the class of all f∈H(U)f \in H(U) for which

∬U∣f(z)∣2dxdy≤1 \iint _{U}\left|f(z)\right|^{2} d x d y \leq 1

Is this a normal family?

?
Problem 2.14.17

Suppose Ω\Omega is a region, fn∈H(Ω)f_{n} \in H(\Omega ) for n=1,2,3,…,fn→fn = 1,2,3, \ldots , f_{n} \rightarrow f uniformly on compact subsets of Ω\Omega, and ff is one-to-one in Ω\Omega. Does it follow that to each compact K⊂ΩK \subset \Omega there corresponds an integer N(K)N(K) such that fnf_{n} is one-to-one on KK for all n>N(K)n > N(K)? Give proof or counterexample.

?
Problem 2.14.18

Suppose Ω\Omega is a simply connected region, z0∈Ωz_{0} \in \Omega, and ff and gg are one-to-one conformal mappings of Ω\Omega onto UU which carry z0z_{0} to 0. What relation exists between ff and gg? Answer the same question if f(z0)=g(z0)=af\left(z_{0}\right) = g\left(z_{0}\right) = a, for some a∈Ua \in U.

?
Problem 2.14.19

Find a homeomorphism of UU onto UU which cannot be extended to a continuous function on Uˉ\bar{U}.

?
Problem 2.14.20

If f∈Sf \in \mathscr {S} (Definition 14.10) and nn is a positive integer, prove that there exists a g∈Sg \in \mathscr {S} such that gn(z)=f(zn)g^{n}(z) = f\left(z^{n}\right) for all z∈Uz \in U.

?
Problem 2.14.21

Find all f∈Sf \in \mathscr {S} such that

?
(a)

f(U)⊃Uf(U) \supset U,

(b)

f(U)⊃Uˉf(U) \supset \bar{U},

(c)

∣a2∣=2\left|a_{2}\right| = 2.

Problem 2.14.22

Suppose ff is a one-to-one conformal mapping of UU onto a square with center at 0, and f(0)=0f(0) = 0. Prove that f(iz)=if(z)f(i z) = i f(z). If f(z)=Σcnznf(z) = \Sigma c_{n} z^{n}, prove that cn=0c_{n} = 0 unless n−1n-1 is a multiple of 4. Generalize this: Replace the square by other simply connected regions with rotational symmetry.

?
Problem 2.14.23

Let Ω\Omega be a bounded region whose boundary consists of two nonintersecting circles. Prove that there is a one-to-one conformal mapping of Ω\Omega onto an annulus. (This is true for every region Ω\Omega such that S2−ΩS^{2}-\Omega has exactly two components, each of which contains more than one point, but this general situation is harder to handle).

?
Problem 2.14.24

Complete the details in the following proof of Theorem 14.22. Suppose 1<R2<R11 < R_{2} < R_{1} and ff is a one-to-one conformal mapping of A(1,R1)A\left(1, R_{1}\right) onto A(1,R2)A\left(1, R_{2}\right). Define f1=ff_{1} = f and fn=f∘fn−1f_{n} = f \circ f_{n-1}. Then a subsequence of {fn}\left\{ f_{n}\right\} converges uniformly on compact subsets of A(1,R1)A\left(1, R_{1}\right) to a function gg. Show that the range of gg cannot contain any nonempty open set (by the three-circle theorem, for instance). On the other hand, show that gg cannot be constant on the circle {z:∣z∣2=R1}\left\{ z:\left|z\right|^{2} = R_{1}\right\}. Hence ff cannot exist.

?
Problem 2.14.25

Here is yet another proof of Theorem 14.22. If ff is as in 14.22, repeated use of the reflection principle extends ff to an entire function such that ∣f(z)∣=1\left|f(z)\right| = 1 whenever ∣z∣=1\left|z\right| = 1. This implies f(z)=αznf(z) = \alpha z^{n}, where ∣α∣=1\left|\alpha \right| = 1 and nn is an integer. Complete the details.

?
Problem 2.14.26

Iteration of Step 2 in the proof of Theorem 14.8 leads to a proof (due to Koebe) of the Riemann mapping theorem which is constructive in the sense that it makes no appeal to the theory of normal families and so does not depend on the existence of some unspecified subsequence. For the final step of the proof it is convenient to assume that Ω\Omega has property (h)(h) of Theorem 13.11. Then any region conformally equivalent to Ω\Omega will satisfy (h)(h). Recall also that (h)(h) implies (j)(j), trivially.

By Step 1 in Theorem 14.8 we may assume, without loss of generality, that 0∈Ω,Ω⊂U0 \in \Omega , \Omega \subset U, and Ω≠U\Omega \neq U. Put Ω=Ω0\Omega = \Omega_{0}. The proof consists in the construction of regions Ω1,Ω2,Ω3,…\Omega_{1}, \Omega_{2}, \Omega_{3}, \ldots and of functions f1,f2,f3,…f_{1}, f_{2}, f_{3}, \ldots, so that fn(Ωn−1)=Ωnf_{n}\left(\Omega_{n-1}\right) = \Omega_{n} and so that the functions fn∘fn−1∘⋯∘f2∘f1f_{n} \circ f_{n-1} \circ \cdots \circ f_{2} \circ f_{1} converge to a conformal mapping of Ω\Omega onto UU.

Complete the details in the following outline.

?
(a)

Suppose Ωn−1\Omega_{n-1} is constructed, let rnr_{n} be the largest number such that D(0;rn)⊂Ωn−1D\left(0 ; r_{n}\right) \subset \Omega_{n-1}, let αn\alpha_{n} be a boundary point of Ωn−1\Omega_{n-1} with ∣αn∣=rn\left|\alpha_{n}\right| = r_{n}, choose βn\beta_{n} so that βn2=−αn\beta_{n}^{2} = -\alpha_{n}, and put

Fn=φ−αn∘S∘φ−βn. F_{n} = \varphi _{-\alpha _{n}} \circ S \circ \varphi _{-\beta _{n}}.

(The notation is as in the proof of Theorem 14.8). Show that FnF_{n} has a holomorphic inverse GnG_{n} in Ωn−1\Omega_{n-1}, and put fn=λnGnf_{n} = \lambda_{n} G_{n}, where λn=∣c∣/c\lambda_{n} = \left|c\right| / c and c=Gn′(0)c = G_{n}^{\prime }(0). (This fnf_{n} is the Koebe mapping associated with Ωn−1\Omega_{n-1}. Note that fnf_{n} is an elementary function. It involves only two linear fractional transformations and a square root).

(b)

Compute that fn′(0)=(1+rn)/2rn>1f_{n}^{\prime }(0) = \left(1+r_{n}\right) / 2 \sqrt{r_{n}} > 1.

(c)

Put ψ0(z)=z\psi_{0}(z) = z and ψn(z)=fn(ψn−1(z))\psi_{n}(z) = f_{n}\left(\psi_{n-1}(z)\right). Show that ψn\psi_{n} is a one-to-one mapping of Ω\Omega onto a region Ωn⊂U\Omega_{n} \subset U, that {ψn′(0)}\left\{ \psi_{n}^{\prime }(0)\right\} is bounded, that

ψn′(0)=∏k=1n1+rk2rk \psi _{n}^{\prime }(0) = \prod _{k = 1}^{n} \frac{1+r_{k}}{2 \sqrt{r_{k}}}

and that therefore rn→1r_{n} \rightarrow 1 as n→∞n \rightarrow \infty.

(d)

Write ψn(z)=zhn(z)\psi_{n}(z) = z h_{n}(z), for z∈Ωz \in \Omega, show that ∣hn∣≤∣hn+1∣\left|h_{n}\right| \leq \left|h_{n+1}\right|, apply Harnack's theorem and Exercise 8 of Chap. 11 to {log⁡∣hn∣}\left\{ \log \left|h_{n}\right|\right\} to prove that {ψn}\left\{ \psi_{n}\right\} converges uniformly on compact subsets of Ω\Omega, and show that lim⁡ψn\lim \psi_{n} is a one-to-one mapping of Ω\Omega onto UU.

Problem 2.14.27

Prove that ∑n=1∞(1−rn)2<∞\sum_{n = 1}^{\infty }\left(1-r_{n}\right)^{2} < \infty, where {rn}\left\{ r_{n}\right\} is the sequence which occurs in Exercise 26. Hint:

1+r2r=1+(1−r)22r \frac{1+r}{2 \sqrt{r}} = 1+\frac{(1-\sqrt{r})^{2}}{2 \sqrt{r}}
?
Problem 2.14.28

Suppose that in Exercise 26 we choose αn∈U−Ωn−1\alpha_{n} \in U-\Omega_{n-1} without insisting that ∣αn∣=rn\left|\alpha_{n}\right| = r_{n}. For example, insist only that

∣αn∣≤1+rn2 \left|\alpha _{n}\right| \leq \frac{1+r_{n}}{2}

Will the resulting sequence {ψn}\left\{ \psi_{n}\right\} still converge to the desired mapping function?

?
Problem 2.14.29

Suppose Ω\Omega is a bounded region, a∈Ω,f∈H(Ω),f(Ω)⊂Ωa \in \Omega , f \in H(\Omega ), f(\Omega ) \subset \Omega, and f(a)=af(a) = a.

?
(a)

Put f1=ff_{1} = f and fn=f∘fn−1f_{n} = f \circ f_{n-1}, compute fn′(a)f_{n}^{\prime }(a), and conclude that ∣f′(a)∣≤1\left|f^{\prime }(a)\right| \leq 1.

(b)

If f′(a)=1f^{\prime }(a) = 1, prove that f(z)=zf(z) = z for all z∈Ωz \in \Omega. Hint: If

f(z)=z+cm(z−a)m+⋯ f(z) = z+c_{m}(z-a)^{m}+\cdots

compute the coefficient of (z−a)m(z-a)^{m} in the expansion of fn(z)f_{n}(z).

(c)

If ∣f′(a)∣=1\left|f^{\prime }(a)\right| = 1, prove that ff is one-to-one and that f(Ω)=Ωf(\Omega ) = \Omega.

Hint: If γ=f′(a)\gamma = f^{\prime }(a), there are integers nk→∞n_{k} \rightarrow \infty such that γnk→1\gamma^{n_{k}} \rightarrow 1 and fnk→gf_{n_{k}} \rightarrow g. Then g′(a)=1g^{\prime }(a) = 1, g(Ω)⊂Ωg(\Omega ) \subset \Omega (by Exercise 20, Chap. 10), hence g(z)=zg(z) = z, by part (b). Use gg to draw the desired conclusions about ff.

Problem 2.14.30

Let Λ\Lambda be the set of all linear fractional transformations. If {α,β,γ,δ}\left\{ \alpha , \beta , \gamma , \delta \right\} is an ordered quadruple of distinct complex numbers, its cross ratio is defined to be

[α,β,γ,δ]=(α−β)(γ−δ)(α−δ)(γ−β) [\alpha , \beta , \gamma , \delta ] = \frac{(\alpha -\beta )(\gamma -\delta )}{(\alpha -\delta )(\gamma -\beta )}

If one of these numbers is ∞\infty, the definition is modified in the obvious way, by continuity. The same applies if α\alpha coincides with β\beta or γ\gamma or δ\delta.

?
(a)

If φ(z)=[z,α,β,γ]\varphi (z) = [z, \alpha , \beta , \gamma ], show that φ∈Λ\varphi \in \Lambda and φ\varphi maps {α,β,γ}\left\{ \alpha , \beta , \gamma \right\} to {0,1,∞}\left\{ 0,1, \infty \right\}.

(b)

Show that the equation [w,a,b,c]=[z,α,β,γ][w, a, b, c] = [z, \alpha , \beta , \gamma ] can be solved in the form w=φ(z)w = \varphi (z); then φ∈Λ\varphi \in \Lambda maps {α,β,γ}\left\{ \alpha , \beta , \gamma \right\} to {a,b,c}\left\{ a, b, c\right\}.

(c)

If φ∈Λ\varphi \in \Lambda, show that

[φ(α),φ(β),φ(γ),φ(δ)]=[α,β,γ,δ] [\varphi (\alpha ), \varphi (\beta ), \varphi (\gamma ), \varphi (\delta )] = [\alpha , \beta , \gamma , \delta ]
(d)

Show that [α,β,γ,δ][\alpha , \beta , \gamma , \delta ] is real if and only if the four points lie on the same circle or straight line.

(e)

Two points zz and z∗z^{*} are said to be symmetric with respect to the circle (or straight line) CC through α,β\alpha , \beta, and γ\gamma if [z∗,α,β,γ]\left[z^{*}, \alpha , \beta , \gamma \right] is the complex conjugate of [z,α,β,γ][z, \alpha , \beta , \gamma ]. If CC is the unit circle, find a simple geometric relation between zz and z∗z^{*}. Do the same if CC is a straight line.

(f)

Suppose zz and z∗z^{*} are symmetric with respect to CC. Show that φ(z)\varphi (z) and φ(z∗)\varphi \left(z^{*}\right) are symmetric with respect to φ(C)\varphi (C), for every φ∈Λ\varphi \in \Lambda.

Problem 2.14.31
?
(a)

Show that Λ\Lambda (see Exercise 30) is a group, with composition as group operation. That is, if φ∈Λ\varphi \in \Lambda and ψ∈Λ\psi \in \Lambda, show that φ∘ψ∈Λ\varphi \circ \psi \in \Lambda and that the inverse φ−1\varphi^{-1} of φ\varphi is in Λ\Lambda. Show that Λ\Lambda is not commutative.

(b)

Show that each member of Λ\Lambda (other than the identity mapping) has either one or two fixed points on S2S^{2}.

(c)

Call two mappings φ\varphi and φ1∈Λ\varphi_{1} \in \Lambda conjugate if there exists a ψ∈Λ\psi \in \Lambda such that φ1=ψ−1∘φ∘ψ\varphi_{1} = \psi^{-1} \circ \varphi \circ \psi. Prove that every φ∈Λ\varphi \in \Lambda with a unique fixed point is conjugate to the mapping z→z+1z \rightarrow z+1. Prove that every φ∈Λ\varphi \in \Lambda with two distinct fixed points is conjugate to the mapping z→αzz \rightarrow \alpha z, where α\alpha is a complex number; to what extent is α\alpha determined by φ\varphi?

(d)

Let α\alpha be a complex number. Show that to every φ∈Λ\varphi \in \Lambda which has α\alpha for its unique fixed point there corresponds a β\beta such that

1φ(z)−α=1z−α+β \frac{1}{\varphi (z)-\alpha } = \frac{1}{z-\alpha }+\beta

Let GαG_{\alpha } be the set of all these φ\varphi, plus the identity transformation. Prove that GαG_{\alpha } is a subgroup of Λ\Lambda and that GαG_{\alpha } is isomorphic to the additive group of all complex numbers.

(e)

Let α\alpha and β\beta be distinct complex numbers, and let Gα,βG_{\alpha , \beta } be the set of all φ∈Λ\varphi \in \Lambda which have α\alpha and β\beta as fixed points. Show that every φ∈Gα,β\varphi \in G_{\alpha , \beta } is given by

φ(z)−αφ(z)−β=γ⋅z−αz−β \frac{\varphi (z)-\alpha }{\varphi (z)-\beta } = \gamma \cdot \frac{z-\alpha }{z-\beta }

where γ\gamma is a complex number. Show that Gα,βG_{\alpha , \beta } is a subgroup of Λ\Lambda which is isomorphic to the multiplicative group of all nonzero complex numbers.

(f)

If φ\varphi is as in (d)(d) or (e)(e), for which circles CC is it true that φ(C)=C\varphi (C) = C? The answer should be in terms of the parameters α,β\alpha , \beta, and γ\gamma.

Remark.
?

In part (b): a fixed point of φ\varphi is a point α\alpha such that φ(α)=α\varphi (\alpha ) = \alpha.

Problem 2.14.32

For z∈Uˉ,z2≠1z \in \bar{U}, z^{2} \neq 1, define

f(z)=exp⁡{ilog⁡1+z1−z} f(z) = \exp \left\{ i \log \frac{1+z}{1-z}\right\}

choosing the branch of log⁡\log that has log⁡1=0\log 1 = 0. Describe f(E)f(E) if EE is

?
(a)

UU,

(b)

the upper half of TT,

(c)

the lower half of TT,

(d)

any circular arc (in UU) from -1 to 1,

(e)

the radius [0,1)[0,1),

(f)

any disc {z:∣z−r∣<1−r},0<r<1\left\{ z:\left|z-r\right| < 1-r\right\} , 0 < r < 1.

(g)

any curve in UU tending to 1.

Problem 2.14.33

If φα\varphi_{\alpha } is as in Definition 12.3, show that

?
(a)

1π∫U∣φα′∣2dm=1\frac{1}{\pi } \int_{U}\left|\varphi_{\alpha }^{\prime }\right|^{2} d m = 1,

(b)

1π∫U∣φα′∣dm=1−∣α∣2∣α∣2log⁡11−∣α∣2\frac{1}{\pi } \int_{U}\left|\varphi_{\alpha }^{\prime }\right| d m = \frac{1-\left|\alpha \right|^{2}}{\left|\alpha \right|^{2}} \log \frac{1}{1-\left|\alpha \right|^{2}}.

Here mm denotes Lebesgue measure in R2\mathbb {R}^{2}.