Argument Principle

Subjects: Complex Analysis
Links: Analytic Functions, Zeros of Analytic Functions, Poles of Analytic Functions, Homology in C

Argument Principle (Local Ver)

Let ΔC be an open ball and a1,,an,b1,,bmΔ. If f is a meromorphic on Δ such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}, then

12πiγf(z)f(z)dz=j=1nkjn(γ,ai)l=1mhln(γ,bl)

for all γΔ{a1,,an,b1,,bm} that is piecewise smooth closed curve

Cor: Let ΔC be an open ball and a1,,an,b1,,bmΔ. If f is a meromorphic on Δ such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}. If γΔ{a1,,an,b1,,bm} that is piecewise smooth closed curve and γ~=fγ, then

n(γ~;0)=12πiγf(z)f(z)dz=j=1nkjn(γ,ai)l=1mhln(γ,bl)

Cor: Let ΔC be an open ball and a1,,an,b1,,bmΔ. If f is a meromorphic on Δ such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}. If γΔ{a1,,an,b1,,bm} that is piecewise smooth closed curve such that n(γ;z) can only take the values of 0 or 1 and γ~=fγ , then

n(γ~;0)=(# of zeros# of poles ) of f trapped by γ

Th (Ahlfors): Let f:ΩC analytic, z0ΔΩ, Δ an open ball, and w0=f(z0). If z0 is a zero of order k of the function g(z)=f(z)w0 for all zΔ, then there exist ε,δ>0 such that wBε(w0){w0} the identity f(z)=w has exactly k solutions in Bδ(z0){z0}. Additionally, we can find δ>0 such that all the soltions of the identity f(z)=w are different from each other.

Cor: Let f:ΩC analytic and UΩ be an open subset. If f is nonconstant, then f[U]C is an open set.

Cor: Let f:ΩC analytic. If f is injective in the region Ω, f(z)0 for all zΩ

Prop: Let f:ΩC analytic. If f injective in the region Ω, then the inverse function f1:f[Ω]Ω is analytic in f[Ω] (is also a region) and additionally (f1)(w)0 for all wΩ

Prop: Let f:ΩC analytic and injective in Ω. If z0Ω and R>0 such that BR(z0)Ω and γr(t)=reit+z0, with t[0,2π] and 0<r<R, then

f1(w)=12πiγrζf(ζ)f(ζ)wdζ

Inverse Function Theorem for Analytic Functions

Let f:ΩC analytic. If z0Ω is such that f(z0)0, then there’s r>0 such that f is injective on Br(z0), f[Br(z0)] is open, f1 is analytic on f[Br(z0)] and

(f1)(w)=1f(f1(w))0

for all wf1[Br(w0)]

Rouché-Gliksberg Theorem

Let f and g be meromorphic functions on the disk ΔC and γΔ be a piecewise smooth closed curve such that n(γ;z)=0 or n(γ;z)=1 for all zΔγ. Let Zf be the number of zeros of f enclosed by γ counted with multiplicities, Pf the number of poles of f enclosed by γ counted with multiplicities, and Zg and Pg be the corresponding ones for the function g. If f and g are such that

|f(z)+g(z)|<|f(z)|+|g(z)|

for all zγ, then

ZfPf=ZgPg

Rouché Theorem

Let f and g be analytic functions on the disk ΔC and γΔ be a piecewise smooth closed curve such that n(γ;z)=0 or n(γ;z)=1 for all zΔγ. Let Zf be the number of zeros of f enclosed by γ counted with multiplicities, Pf the number of poles of f enclosed by γ counted with multiplicities, and Zg and Pg be the corresponding ones for the function g. If f and g are such that

|f(z)g(z)|<|g(z)|

for all zγ, then

Zf=Zg

Hurwitz's Theorem

Let fn,f:ΔC analytics on the disk Δ and γΔ be a piecewise smooth closed curve such that n(γ;z)=0 or n(γ;z)=1 for all zΔγ. If the the sequence fn converges uniformly to f over γ and f(z)0 for all zγ, then there exists NN such that for all nN the functions fn and f have the same amount of zeros enclosed by γ

Prop: Let fn:ΩC be a sequence of analytic functions such that it converges uniformly to f over any compact set KΩ. If fn(z)0 for all nN and for all zΩ, then f is the constant 0 or f(z)0 for all zΩ

Argument Principle

Let ΩC be an region and a1,,an,b1,,bmΩ. If f is a meromorphic on Ω such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}, then

12πiγf(z)f(z)dz=j=1nkjn(γ,ai)l=1mhln(γ,bl)

for any cycle γΩ{a1,,an,b1,,bm} that is γ0(modΩ). This can be done using the Residue Theorem

Cor: Let ΩC be an open ball and a1,,an,b1,,bmΩ. If f is a meromorphic on Ω such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}. If γΩ{a1,,an,b1,,bm} that cycle that γ0(modΩ) and γ~=fγ, then

n(γ~;0)=12πiγf(z)f(z)dz=j=1nkjn(γ,ai)l=1mhln(γ,bl)

Cor: Let ΩC be an open ball and a1,,an,b1,,bmΩ. If f is a meromorphic on Ω such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}. If γΩ{a1,,an,b1,,bm} that is cycle such that γ0(modΩ) and n(γ;z) can only take the values of 0 or 1 and γ~=fγ , then

n(γ~;0)=(# of zeros# of poles ) of f trapped by γ

We can extend this to get that:

Let ΩC be an region and a1,,an,b1,,bmΩ. If f is a meromorphic on Ω such that aj is a zero of order kjN+, for each j{1,,n} and bl is a pole of f of order hl for l{1,,m}, and let g be analyctic on Ω

12πiγg(z)f(z)f(z)dz=j=1nkjn(γ,aj)g(aj)l=1mhln(γ,bl)g(bl)

for any cycle γΩ{a1,,an,b1,,bm} that is γ0(modΩ)