Chapter 2.16

Analytic Continuation

[15/0%]
Show
LaTeX
Problem 2.16.1

Suppose f(z)=Σanzn,an≥0f(z) = \Sigma a_{n} z^{n}, a_{n} \geq 0, and the radius of convergence of the series is 1. Prove that ff has a singularity at z=1z = 1. Hint: Expand ff in powers of z−12z-\frac{1}{2}. If 1 were a regular point of ff, the new series would converge at some x>1x > 1. What would this imply about the original series?

?
Problem 2.16.2

Suppose (f,D)(f, D) and (g,D)(g, D) are function elements, PP is a polynomial in two variables, and P(f,g)=0P(f, g) = 0 in DD. Suppose ff and gg can be analytically continued along a curve γ\gamma, to f1f_{1} and g1g_{1}. Prove that P(f1,g1)=0P\left(f_{1}, g_{1}\right) = 0. Extend this to more than two functions. Is there such a theorem for some class of functions PP which is larger than the polynomials?

?
Problem 2.16.3

Suppose Ω\Omega is a simply connected region, and uu is a real harmonic function in Ω\Omega. Prove that there exists an f∈H(Ω)f \in H(\Omega ) such that u=Re(f)u = \mathrm{Re}\left(f\right). Show that this fails in every region which is not simply connected.

?
Problem 2.16.4

Suppose XX is the closed unit square in the plane, ff is a continuous complex function on XX, and ff has no zero in XX. Prove that there is a continuous function gg on XX such that f=egf = e^{g}. For what class of spaces XX (other than the above square) is this also true?

?
Problem 2.16.5

Prove that the transformations z→z+1z \rightarrow z+1 and z→−1/zz \rightarrow -1 / z generate the full modular group GG. Let RR consist of all z=x+iyz = x+i y such that ∣x∣<12,y>0\left|x\right| < \frac{1}{2}, y > 0, and ∣z∣>1\left|z\right| > 1, plus those limit points which have x≤0x \leq 0. Prove that RR is a fundamental domain of GG.

?
Problem 2.16.6

Prove that GG is also generated by the transformations φ\varphi and ψ\psi, where

φ(z)=−1z,ψ(z)=z−1z \varphi (z) = -\frac{1}{z}, \quad \psi (z) = \frac{z-1}{z}

Show that φ\varphi has period 2,ψ2, \psi has period 3.

?
Problem 2.16.7

Find the relation between composition of linear fractional transformations and matrix multiplication. Try to use this to construct an algebraic proof of Theorem 16.19(c) or of the first part of Exercise 5.

?
Problem 2.16.8

Let EE be a compact set on the real axis, of positive Lebesgue measure, let Ω\Omega be the complement of EE, relative to the plane, and define

f(z)=∫Edtt−z(z∈Ω) f(z) = \int _{E} \frac{d t}{t-z} \quad (z \in \Omega )

Answer the following questions:

?
(a)

Is ff constant?

(b)

Can ff be extended to an entire function?

(c)

Does lim⁡zf(z)\lim z f(z) exist as z→∞z \rightarrow \infty? If so, what is it?

(d)

Does ff have a holomorphic square root in Ω\Omega?

(e)

Is the real part of ff bounded in Ω\Omega?

(f)

Is the imaginary part of ff bounded in Ω\Omega? If "yes" in (e) or (f), give a bound.

(g)

What is ∫γf(z)dz\int_{\gamma } f(z) d z if γ\gamma is a positively oriented circle which has EE in its interior?

(h)

Does there exist a bounded holomorphic function φ\varphi in Ω\Omega which is not constant?

Problem 2.16.9

Check your answers in Exercise 8 against the special case

E=[−1,1] E = [-1,1]
?
Problem 2.16.10

Call a compact set EE in the plane removable if there are no nonconstant bounded holomorphic functions in the complement of EE.

?
(a)

Prove that every countable compact set is removable.

(b)

If EE is a compact subset of the real axis, and m(E)=0m(E) = 0, prove that EE is removable. Hint: EE can be surrounded by curves of arbitrarily small total length. Apply Cauchy's formula, as in Exercise 25, Chap 10.

(c)

Suppose EE is removable, Ω\Omega is a region, E⊂Ω,f∈H(Ω−E)E \subset \Omega , f \in H(\Omega -E), and ff is bounded. Prove that ff can be extended to a holomorphic function in Ω\Omega.

(d)

Formulate and prove an analogue of part (b) for sets EE which are not necessarily on the real axis.

(e)

Prove that no compact connected subset of the plane (with more than one point) is removable.

Problem 2.16.11

For each positive number α\alpha, let Γα\Gamma_{\alpha } be the path with parameter interval (−∞,∞)(-\infty , \infty ) defined by

Γα(t)={−t−πi(−∞<t≤−α)α+πitα(−α≤t≤α)t+πi(α≤t<∞) \Gamma _{\alpha }(t) = \begin{cases} -t-\pi i & (-\infty < t \leq -\alpha ) \\ \alpha +\frac{\pi i t}{\alpha } & (-\alpha \leq t \leq \alpha ) \\ t+\pi i & (\alpha \leq t < \infty )\end{cases}

Let Ωα\Omega_{\alpha } be the component of the complement of Γα∗\Gamma_{\alpha }^{*} which contains the origin, and define

fα(z)=12πi∫Γαexp⁡(ew)w−zdw(z∈Ωα) f_{\alpha }(z) = \frac{1}{2 \pi i} \int _{\Gamma _{\alpha }} \frac{\exp \left(e^{w}\right)}{w-z} d w \quad \left(z \in \Omega _{\alpha }\right)

Prove that fβf_{\beta } is an analytic continuation of fαf_{\alpha } if α<β\alpha < \beta. Prove that therefore there is an entire function ff whose restriction to Ωα\Omega_{\alpha } is fαf_{\alpha }. Prove that

lim⁡r→∞f(reiθ)=0 \lim _{r \rightarrow \infty } f\left(r e^{i \theta }\right) = 0

for every eiθ≠1e^{i \theta } \neq 1. (Here rr is positive and θ\theta is real, as usual). Prove that ff is not constant. Hint: Look at f(r)f(r). If

g=fexp⁡(−f) g = f \exp (-f)

prove that

lim⁡r→∞g(reiθ)=0 \lim _{r \rightarrow \infty } g\left(r e^{i \theta }\right) = 0

for every eiθe^{i \theta }. Show that there exists an entire function hh such that

lim⁡n→∞h(nz)={1 if z=00 if z≠0 \lim _{n \rightarrow \infty } h(n z) = \begin{cases} 1 & \text{ if } z = 0 \\ 0 & \text{ if } z \neq 0\end{cases}
?
Problem 2.16.12

Suppose

f(z)=∑k=1∞(z−z22)3k=∑n=1∞anzn f(z) = \sum _{k = 1}^{\infty }\left(\frac{z-z^{2}}{2}\right)^{3^{k}} = \sum _{n = 1}^{\infty } a_{n} z^{n}

Find the regions in which the two series converge. Show that this illustrates Theorem 16.5. Find the singular point of ff which is nearest to the origin.

?
Problem 2.16.13

Let Ω={z:12<∣z∣<2}\Omega = \left\{ z: \frac{1}{2} < \left|z\right| < 2\right\}. For n=1,2,3,…n = 1,2,3, \ldots let XnX_{n} be the set of all f∈H(Ω)f \in H(\Omega ) that are nnth derivatives of some g∈H(Ω)g \in H(\Omega ).

?
(a)

Show that f∈X1f \in X_{1} if and only if ∫γf(z)dz=0\int_{\gamma } f(z) d z = 0, where γ\gamma is the positively oriented unit circle.

(b)

Show that f∈Xnf \in X_{n} for every nn if and only if ff extends to a holomorphic function in D(0;2)D(0 ; 2).

Remark.
?

In other words, XnX_{n} is the range of the differential operator DnD^{n} with domain H(Ω)H(\Omega ).

Problem 2.16.14

Suppose Ω\Omega is a region, p∈Ω,R<∞p \in \Omega , R < \infty. Let F\mathscr {F} be the class of all f∈H(Ω)f \in H(\Omega ) such that ∣f(p)∣≤R\left|f(p)\right| \leq R and f(Ω)f(\Omega ) contains neither 0 nor 1. Prove that F\mathscr {F} is a normal family.

?
Problem 2.16.15

Show that Theorem 16.2 leads to a very simple proof of the special case of the monodromy theorem (16.15) in which Ω\Omega and DD are concentric discs. Combine this special case with the Riemann mapping theorem to prove Theorem 16.15 in the generality in which it is stated.

?