1.8 Fourier Series1.10 Functions of a Complex Variable

§1.9 Calculus of a Complex Variable

Contents

§1.9(i) Complex Numbers

1.9.1z=x+iy,x,y\in\Real.

Real and Imaginary Parts

1.9.2
\realpart{z}=x,
\imagpart{z}=y.

Polar Representation

1.9.3
x=r\mathop{\cos\/}\nolimits\theta,
y=r\mathop{\sin\/}\nolimits\theta,

where

1.9.4r=(x^{2}+y^{2})^{{1/2}},

and when z\neq 0,

1.9.5\theta=\omega,\;\;\pi-\omega,\;\;-\pi+\omega,\mbox{ or }-\omega,

according as z lies in the 1st, 2nd, 3rd, or 4th quadrants. Here

1.9.6\omega=\mathop{\mathrm{arctan}\/}\nolimits\!\left(|y/x|\right)\in\left[0,\tfrac{1}{2}\pi\right].

Modulus and Phase

1.9.7
|z|=r,
\mathop{\mathrm{ph}\/}\nolimits z=\theta+2n\pi,n\in\Integer.

The principal value of \mathop{\mathrm{ph}\/}\nolimits z corresponds to n=0, that is, -\pi\leq\mathop{\mathrm{ph}\/}\nolimits z\leq\pi. It is single-valued on \Complex\setminus\{ 0\}, except on the interval (-\infty,0) where it is discontinuous and two-valued. Unless indicated otherwise, these principal values are assumed throughout the DLMF. (However, if we require a principal value to be single-valued, then we can restrict -\pi<\mathop{\mathrm{ph}\/}\nolimits z\leq\pi.)

1.9.8
|\realpart{z}|\leq|z|,
|\imagpart{z}|\leq|z|,
1.9.9z=re^{{i\theta}},

where

1.9.10e^{{i\theta}}=\mathop{\cos\/}\nolimits\theta+i\mathop{\sin\/}\nolimits\theta;

see §4.14.

Complex Conjugate

1.9.11\conj{z}=x-iy,
1.9.12|\conj{z}|=|z|,
1.9.13\mathop{\mathrm{ph}\/}\nolimits\conj{z}=-\mathop{\mathrm{ph}\/}\nolimits z.

Arithmetic Operations

If z_{1}=x_{1}+iy_{1}, z_{2}=x_{2}+iy_{2}, then

1.9.14z_{1}\pm z_{2}=x_{1}\pm x_{2}+i(y_{1}\pm y_{2}),
1.9.15z_{1}z_{2}=x_{1}x_{2}-y_{1}y_{2}+i(x_{1}y_{2}+x_{2}y_{1}),
1.9.16\frac{z_{1}}{z_{2}}=\frac{z_{1}\conj{z}_{2}}{|z_{2}|^{2}}=\frac{x_{1}x_{2}+y_{1}y_{2}+i(x_{2}y_{1}-x_{1}y_{2})}{x_{2}^{2}+y_{2}^{2}},

provided that z_{2}\neq 0. Also,

1.9.17|z_{1}z_{2}|=|z_{1}|\;|z_{2}|,
1.9.18\mathop{\mathrm{ph}\/}\nolimits\!\left(z_{1}z_{2}\right)=\mathop{\mathrm{ph}\/}\nolimits z_{1}+\mathop{\mathrm{ph}\/}\nolimits z_{2},
1.9.19\left|\frac{z_{1}}{z_{2}}\right|=\frac{|z_{1}|}{|z_{2}|},
1.9.20\mathop{\mathrm{ph}\/}\nolimits\frac{z_{1}}{z_{2}}=\mathop{\mathrm{ph}\/}\nolimits z_{1}-\mathop{\mathrm{ph}\/}\nolimits z_{2}.

Equations (1.9.18) and (1.9.20) hold for general values of the phases, but not necessarily for the principal values.

Powers

1.9.21z^{n}=\left(x^{n}-\binom{n}{2}x^{{n-2}}y^{2}+\binom{n}{4}x^{{n-4}}y^{4}-\cdots\right)+i\left(\binom{n}{1}x^{{n-1}}y-\binom{n}{3}x^{{n-3}}y^{3}+\cdots\right),n=1,2,\dots.

Triangle Inequality

1.9.23\left|\left|z_{1}\right|-\left|z_{2}\right|\right|\leq\left|z_{1}+z_{2}\right|\leq\left|z_{1}\right|+\left|z_{2}\right|.

§1.9(ii) Continuity, Point Sets, and Differentiation

Continuity

A function f(z) is continuous at a point z_{0} if \lim\limits _{{z\to z_{0}}}f(z)=f(z_{0}). That is, given any positive number \epsilon, however small, we can find a positive number \delta such that |f(z)-f(z_{0})|<\epsilon for all z in the open disk |z-z_{0}|<\delta.

A function of two complex variables f(z,w) is continuous at (z_{0},w_{0}) if \lim\limits _{{(z,w)\to(z_{0},w_{0})}}f(z,w)=f(z_{0},w_{0}); compare (1.5.1) and (1.5.2).

Point Sets in \Complex

A neighborhood of a point z_{0} is a disk \left|z-z_{0}\right|<\delta. An open set in \Complex is one in which each point has a neighborhood that is contained in the set.

A point z_{0} is a limit point (limiting point or accumulation point) of a set of points S in \Complex (or \Complex\cup\infty) if every neighborhood of z_{0} contains a point of S distinct from z_{0}. (z_{0} may or may not belong to S.) As a consequence, every neighborhood of a limit point of S contains an infinite number of points of S. Also, the union of S and its limit points is the closure of S.

A domain D, say, is an open set in \Complex that is connected, that is, any two points can be joined by a polygonal arc (a finite chain of straight-line segments) lying in the set. Any point whose neighborhoods always contain members and nonmembers of D is a boundary point of D. When its boundary points are added the domain is said to be closed, but unless specified otherwise a domain is assumed to be open.

A region is an open domain together with none, some, or all of its boundary points. Points of a region that are not boundary points are called interior points.

A function f(z) is continuous on a region R if for each point z_{0} in R and any given number \epsilon (>0) we can find a neighborhood of z_{0} such that \left|f(z)-f(z_{0})\right|<\epsilon for all points z in the intersection of the neighborhood with R.

Differentiation

A function f(z) is differentiable at a point z if the following limit exists:

1.9.24f^{{\prime}}(z)=\frac{df}{dz}=\lim _{{h\to 0}}\frac{f(z+h)-f(z)}{h}.

Differentiability automatically implies continuity.

Cauchy–Riemann Equations

If f^{{\prime}}(z) exists at z=x+iy and f(z)=u(x,y)+iv(x,y), then

1.9.25
\frac{\partial u}{\partial x}=\frac{\partial v}{\partial y},
\frac{\partial u}{\partial y}=-\frac{\partial v}{\partial x}

at (x,y).

Conversely, if at a given point (x,y) the partial derivatives \ifrac{\partial u}{\partial x}, \ifrac{\partial u}{\partial y}, \ifrac{\partial v}{\partial x}, and \ifrac{\partial v}{\partial y} exist, are continuous, and satisfy (1.9.25), then f(z) is differentiable at z=x+iy.

Analyticity

A function f(z) is said to be analytic (holomorphic) at z=z_{0} if it is differentiable in a neighborhood of z_{0}.

A function f(z) is analytic in a domain D if it is analytic at each point of D. A function analytic at every point of \Complex is said to be entire.

If f(z) is analytic in an open domain D, then each of its derivatives f^{{\prime}}(z), f^{{\prime\prime}}(z), \dots exists and is analytic in D.

Harmonic Functions

If f(z)=u(x,y)+iv(x,y) is analytic in an open domain D, then u and v are harmonic in D, that is,

1.9.26\frac{{\partial}^{2}u}{{\partial x}^{2}}+\frac{{\partial}^{2}u}{{\partial y}^{2}}=\frac{{\partial}^{2}v}{{\partial x}^{2}}+\frac{{\partial}^{2}v}{{\partial y}^{2}}=0,

or in polar form ((1.9.3)) u and v satisfy

1.9.27\frac{{\partial}^{2}u}{{\partial r}^{2}}+\frac{1}{r}\frac{\partial u}{\partial r}+\frac{1}{r^{2}}\frac{{\partial}^{2}u}{{\partial\theta}^{2}}=0

at all points of D.

§1.9(iii) Integration

An arc C is given by z(t)=x(t)+iy(t), a\leq t\leq b, where x and y are continuously differentiable. If x(t) and y(t) are continuous and x^{{\prime}}(t) and y^{{\prime}}(t) are piecewise continuous, then z(t) defines a contour.

A contour is simple if it contains no multiple points, that is, for every pair of distinct values t_{1},t_{2} of t, z(t_{1})\neq z(t_{2}). A simple closed contour is a simple contour, except that z(a)=z(b).

Next,

1.9.28\int _{C}f(z)dz=\int _{a}^{b}f(z(t))(x^{{\prime}}(t)+iy^{{\prime}}(t))dt,

for a contour C and f(z(t)) continuous, a\leq t\leq b. If f(z(t_{0}))=\infty, a\leq t_{0}\leq b, then the integral is defined analogously to the infinite integrals in §1.4(v). Similarly when a=-\infty or b=+\infty.

Jordan Curve Theorem

Any simple closed contour C divides \Complex into two open domains that have C as common boundary. One of these domains is bounded and is called the interior domain of C; the other is unbounded and is called the exterior domain of C.

Cauchy’s Theorem

If f(z) is continuous within and on a simple closed contour C and analytic within C, then

1.9.29\int _{C}f(z)dz=0.

Cauchy’s Integral Formula

If f(z) is continuous within and on a simple closed contour C and analytic within C, and if z_{0} is a point within C, then

1.9.30f(z_{0})=\frac{1}{2\pi i}\int _{C}\frac{f(z)}{z-z_{0}}dz,

and

1.9.31f^{{(n)}}(z_{0})=\frac{n!}{2\pi i}\int _{C}\frac{f(z)}{(z-z_{0})^{{n+1}}}dz,n=1,2,3,\dots,

provided that in both cases C is described in the positive rotational (anticlockwise) sense.

Liouville’s Theorem

Any bounded entire function is a constant.

Winding Number

If C is a closed contour, and z_{0}\not\in C, then

1.9.32\frac{1}{2\pi i}\int _{C}\frac{1}{z-z_{0}}dz=\mathcal{N}(C,z_{0}),

where \mathcal{N}(C,z_{0}) is an integer called the winding number of C with respect to z_{0}. If C is simple and oriented in the positive rotational sense, then \mathcal{N}(C,z_{0}) is 1 or 0 depending whether z_{0} is inside or outside C.

Poisson Integral

If h(w) is continuous on |w|=R, then with z=re^{{i\theta}}

1.9.34u(re^{{i\theta}})=\frac{1}{2\pi}\int^{{2\pi}}_{0}\frac{(R^{2}-r^{2})h(Re^{{i\phi}})d\phi}{R^{2}-2Rr\mathop{\cos\/}\nolimits\!\left(\phi-\theta\right)+r^{2}}

is harmonic in |z|<R. Also with \left|w\right|=R, \lim\limits _{{z\to w}}u(z)=h(w) as z\to w within |z|<R.

§1.9(iv) Conformal Mapping

The extended complex plane, \Complex\,\cup\,\{\infty\}, consists of the points of the complex plane \Complex together with an ideal point \infty called the point at infinity. A system of open disks around infinity is given by

1.9.35S_{r}=\{ z\mid|z|>1/r\}\cup\{\infty\},0<r<\infty.

Each S_{r} is a neighborhood of \infty. Also,

1.9.36\infty\pm z=z\pm\infty=\infty,
1.9.37\infty\cdot z=z\cdot\infty=\infty,z\not=0,
1.9.38z/\infty=0,
1.9.39z/0=\infty,z\neq 0.

A function f(z) is analytic at \infty if g(z)=f(1/z) is analytic at z=0, and we set f^{{\prime}}(\infty)=g^{{\prime}}(0).

Conformal Transformation

Suppose f(z) is analytic in a domain D and C_{1},C_{2} are two arcs in D passing through z_{0}. Let C^{{\prime}}_{1},C^{{\prime}}_{2} be the images of C_{1} and C_{2} under the mapping w=f(z). The angle between C_{1} and C_{2} at z_{0} is the angle between the tangents to the two arcs at z_{0}, that is, the difference of the signed angles that the tangents make with the positive direction of the real axis. If f^{{\prime}}(z_{0})\not=0, then the angle between C_{1} and C_{2} equals the angle between C^{{\prime}}_{1} and C^{{\prime}}_{2} both in magnitude and sense. We then say that the mapping w=f(z) is conformal (angle-preserving) at z_{0}.

The linear transformation f(z)=az+b, a\not=0, has f^{{\prime}}(z)=a and w=f(z) maps \Complex conformally onto \Complex.

Bilinear Transformation

1.9.40w=f(z)=\frac{az+b}{cz+d},ad-bc\not=0, c\not=0.
1.9.41
f(-d/c)=\infty,
f(\infty)=a/c.
1.9.42f^{{\prime}}(z)=\frac{ad-bc}{(cz+d)^{2}},z\not=-d/c.
1.9.43f^{{\prime}}(\infty)=\frac{bc-ad}{c^{2}}.
1.9.44z=\frac{dw-b}{-cw+a}.

The transformation (1.9.40) is a one-to-one conformal mapping of \Complex\,\cup\,\{\infty\} onto itself.

The cross ratio of z_{1},z_{2},z_{3},z_{4}\in\Complex\cup\{\infty\} is defined by

1.9.45\frac{(z_{1}-z_{2})(z_{3}-z_{4})}{(z_{1}-z_{4})(z_{3}-z_{2})},

or its limiting form, and is invariant under bilinear transformations.

Other names for the bilinear transformation are fractional linear transformation, homographic transformation, and Möbius transformation.

§1.9(v) Infinite Sequences and Series

A sequence \{ z_{n}\} converges to z if \lim\limits _{{n\to\infty}}z_{n}=z. For z_{n}=x_{n}+iy_{n}, the sequence \{ z_{n}\} converges iff the sequences \{ x_{n}\} and \{ y_{n}\} separately converge. A series \sum^{\infty}_{{n=0}}z_{n} converges if the sequence s_{n}=\sum^{n}_{{k=0}}z_{k} converges. The series is divergent if s_{n} does not converge. The series converges absolutely if \sum^{\infty}_{{n=0}}|z_{n}| converges. A series \sum^{\infty}_{{n=0}}z_{n} converges (diverges) absolutely when \lim\limits _{{n\to\infty}}|z_{n}|^{{1/n}}<1 (>1), or when \lim\limits _{{n\to\infty}}\left|\ifrac{z_{{n+1}}}{z_{n}}\right|<1 (>1). Absolutely convergent series are also convergent.

Let \{ f_{n}(z)\} be a sequence of functions defined on a set S. This sequence converges pointwise to a function f(z) if

1.9.46f(z)=\lim _{{n\to\infty}}f_{n}(z)

for each z\in S. The sequence converges uniformly on S, if for every \epsilon>0 there exists an integer N, independent of z, such that

1.9.47|f_{n}(z)-f(z)|<\epsilon

for all z\in S and n\geq N.

A series \sum^{\infty}_{{n=0}}f_{n}(z) converges uniformly on S, if the sequence s_{n}(z)=\sum^{n}_{{k=0}}f_{k}(z) converges uniformly on S.

Weierstrass M-test

Suppose \{ M_{n}\} is a sequence of real numbers such that \sum^{\infty}_{{n=0}}M_{n} converges and |f_{n}(z)|\leq M_{n} for all z\in S and all n\geq 0. Then the series \sum^{\infty}_{{n=0}}f_{n}(z) converges uniformly on S.

A doubly-infinite series \sum^{\infty}_{{n=-\infty}}f_{n}(z) converges (uniformly) on S iff each of the series \sum^{\infty}_{{n=0}}f_{n}(z) and \sum^{\infty}_{{n=1}}f_{{-n}}(z) converges (uniformly) on S.

§1.9(vi) Power Series

For a series \sum^{\infty}_{{n=0}}a_{n}(z-z_{0})^{n} there is a number R, 0\leq R\leq\infty, such that the series converges for all z in |z-z_{0}|<R and diverges for z in |z-z_{0}|>R. The circle |z-z_{0}|=R is called the circle of convergence of the series, and R is the radius of convergence. Inside the circle the sum of the series is an analytic function f(z). For z in |z-z_{0}|\leq\rho (<R), the convergence is absolute and uniform. Moreover,

1.9.48a_{n}=\frac{f^{{(n)}}(z_{0})}{n!},

and

1.9.49R=\liminf _{{n\to\infty}}|a_{n}|^{{-1/n}}.

For the converse of this result see §1.10(i).

Operations

When \sum a_{n}z^{n} and \sum b_{n}z^{n} both converge

1.9.50\sum^{\infty}_{{n=0}}(a_{n}\pm b_{n})z^{n}=\sum^{\infty}_{{n=0}}a_{n}z^{n}\pm\sum^{\infty}_{{n=0}}b_{n}z^{n},

and

1.9.51\left(\sum^{\infty}_{{n=0}}a_{n}z^{n}\right)\left(\sum^{\infty}_{{n=0}}b_{n}z^{n}\right)=\sum^{\infty}_{{n=0}}c_{n}z^{n},

where

1.9.52c_{n}=\sum^{n}_{{k=0}}a_{k}b_{{n-k}}.

Next, let

1.9.53f(z)=a_{0}+a_{1}z+a_{2}z^{2}+\cdots,a_{0}\not=0.

Then the expansions (1.9.54), (1.9.57), and (1.9.60) hold for all sufficiently small |z|.

1.9.54\frac{1}{f(z)}=b_{0}+b_{1}z+b_{2}z^{2}+\cdots,

where

1.9.55
b_{0}=1/a_{0},
b_{1}=-a_{1}/a_{0}^{2},
b_{2}=(a_{1}^{2}-a_{0}a_{2})/a_{0}^{3},
1.9.56b_{n}=-(a_{1}b_{{n-1}}+a_{2}b_{{n-2}}+\dots+a_{n}b_{0})/a_{0},n\geq 1.

With a_{0}=1,

1.9.57\mathop{\ln\/}\nolimits f(z)=q_{1}z+q_{2}z^{2}+q_{3}z^{3}+\cdots,

(principal value), where

1.9.58
q_{1}=a_{1},
q_{2}=(2a_{2}-a_{1}^{2})/2,
q_{3}=(3a_{3}-3a_{1}a_{2}+a_{1}^{3})/3,

and

1.9.59q_{n}=(na_{n}-(n-1)a_{1}q_{{n-1}}-(n-2)a_{2}q_{{n-2}}-\cdots-a_{{n-1}}q_{1})/n,n\geq 2.

Also,

1.9.60(f(z))^{\nu}=p_{0}+p_{1}z+p_{2}z^{2}+\cdots,

(principal value), where \nu\in\Complex,

1.9.61
p_{0}=1,
p_{1}=\nu a_{1},
p_{2}=\nu((\nu-1)a_{1}^{2}+2a_{2})/2,

and

1.9.62p_{n}=((\nu-n+1)a_{1}p_{{n-1}}+(2\nu-n+2)a_{2}p_{{n-2}}+\dots+((n-1)\nu-1)a_{{n-1}}p_{1}+n\nu a_{n})/n,n\geq 1.

For the definitions of the principal values of \mathop{\ln\/}\nolimits f(z) and (f(z))^{\nu} see §§4.2(i) and 4.2(iv).

Lastly, a power series can be differentiated any number of times within its circle of convergence:

1.9.63f^{{(m)}}(z)=\sum _{{n=0}}^{\infty}\left(n+1\right)_{{m}}a_{{n+m}}(z-z_{0})^{n},\left|z-z_{0}\right|<R, m=0,1,2,\dots.

§1.9(vii) Inversion of Limits

Double Sequences and Series

A set of complex numbers \{ z_{{m,n}}\} where m and n take all positive integer values is called a double sequence. It converges to z if for every \epsilon>0, there is an integer N such that

1.9.64|z_{{m,n}}-z|<\epsilon

for all m,n\geq N. Suppose \{ z_{{m,n}}\} converges to z and the repeated limits

1.9.65
\lim _{{m\to\infty}}\left(\lim _{{n\to\infty}}z_{{m,n}}\right),
\lim _{{n\to\infty}}\left(\lim _{{m\to\infty}}z_{{m,n}}\right)

exist. Then both repeated limits equal z.

A double series is the limit of the double sequence

1.9.66z_{{p,q}}=\sum^{p}_{{m=0}}\sum^{q}_{{n=0}}\zeta _{{m,n}}.

If the limit exists, then the double series is convergent; otherwise it is divergent. The double series is absolutely convergent if it is convergent when \zeta _{{m,n}} is replaced by |\zeta _{{m,n}}|.

If a double series is absolutely convergent, then it is also convergent and its sum is given by either of the repeated sums

1.9.67
\sum^{\infty}_{{m=0}}\left(\sum^{\infty}_{{n=0}}\zeta _{{m,n}}\right),
\sum^{\infty}_{{n=0}}\left(\sum^{\infty}_{{m=0}}\zeta _{{m,n}}\right).

Term-by-Term Integration

Suppose the series \sum^{\infty}_{{n=0}}f_{n}(z), where f_{n}(z) is continuous, converges uniformly on every compact set of a domain D, that is, every closed and bounded set in D. Then

1.9.68\int _{C}\sum^{\infty}_{{n=0}}f_{n}(z)dz=\sum^{\infty}_{{n=0}}\int _{C}f_{n}(z)dz

for any finite contour C in D.

Dominated Convergence Theorem

Let (a,b) be a finite or infinite interval, and f_{0}(t),f_{1}(t),\dots be real or complex continuous functions, t\in(a,b). Suppose \sum^{\infty}_{{n=0}}f_{n}(t) converges uniformly in any compact interval in (a,b), and at least one of the following two conditions is satisfied:

1.9.69\int^{b}_{a}\sum^{\infty}_{{n=0}}|f_{n}(t)|dt<\infty,
1.9.70\sum^{\infty}_{{n=0}}\int^{b}_{a}|f_{n}(t)|dt<\infty.

Then

1.9.71\int^{b}_{a}\sum^{\infty}_{{n=0}}f_{n}(t)dt=\sum^{\infty}_{{n=0}}\int^{b}_{a}f_{n}(t)dt.