In this lecture, first we shall introduce a complex-valued function of a complex variable, and then for such a function define the concept of limit and continuity at a point.
LetSbe a set of complex numbers. Acomplex function(complex-valued of a complex variable)f defined onSis a rule that assigns to eachz=x+iy in S a unique complex numberw=u+iv and written as f :S →C.The numberwis called the value off atzand is denoted byf(z); i.e.,w=f(z).
The set S is called the domain of f, the set W = {f(z) : z ∈ S}, often denoted asf(S),is called the rangeorimage off,andf is said to mapS ontoW.The functionw=f(z) is said to be fromS intoW if the range of S underf is a subset ofW.When a function is given by a formula and the domain is not specified, the domain is taken to be the largest set on which the formula is defined. A function f is calledone-to-one (orunivalent, or injective) on a setSif the equationf(z1) =f(z2),wherez1andz2are inS, implies thatz1 =z2. The functionf(z) = iz is one-to-one, butf(z) =z2 is not one-to-one sincef(i) =f(−i) =−1.A one-to-one and onto function is called bijective. We shall also consider multi-valued functions: amulti- valued function is a rule that assigns a finite or infinite non-empty subset ofC for each element of its domain S.In Lecture 2, we have already seen that the functionf(z) = argz is multi-valued.
As every complex number z is characterized by a pair of real numbers xand y, a complex functionf of the complex variablez can be specified by two real functionsu=u(x, y) andv=v(x, y). It is customary to write w=f(z) =u(x, y)+iv(x, y).The functionsuandv,respectively, are called the real and imaginary parts off. The common domain of the functionsu andv corresponds to the domain of the functionf.
Example 5.1.
For the functionw=f(z) = 3z2+ 7z,we have f(x+iy) = 3(x+iy)2+ 7(x+iy) = (3x2−3y2+ 7x) +i(6xy+ 7y), and henceu= 3x2−3y2+ 7xandv= 6xy+ 7y.Similarly, for the function w=f(z) =|z|2,we findf(x+iy) = |x+iy|2 = x2+y2,
and henceu=x2+y2 andv= 0.Thus, this function is a real-valued func- tion of a complex variable. Clearly, the domain of both of these functions is
R.P. Agarwal et al., An Introduction to Complex Analysis,
DOI 10.1007/978-1-4614-0195-7_5, © Springer Science+Business Media, LLC 2011 28
Complex Functions 29 C. For the functionw=f(z) =z/|z|,the domain is C\{0},and its range is|z|= 1.
Example 5.2.
Thecomplex exponential function f(z) =ez is defined by the formula (3.1). Clearly, for this function,u=excosyandv =exsiny, which are defined for all (x, y)∈IR2. Thus, for the functionezthe domain isC.The exponential function provides a basic tool for the application of complex variables to electrical circuits, control systems, wave propagation, and time-invariant physical systems.Recall that a vector-valued function of two real variables F(x, y) = (P(x, y), Q(x, y)) is also called a two-dimensional vector filed. Using the standard orthogonal unit basis vectors iand j,we can express this vector field as F(x, y) =P(x, y)i+Q(x, y)j. There is a natural way to represent this vector field with a complex function f(z). In fact, we can use the functions P andQas the real and imaginary parts of f,in which case we say that the complex function f(z) = P(x, y) +iQ(x, y) is the complex representationof the vector fieldF(x, y) =P(x, y)i+Q(x, y)j.Conversely, any complex function f(z) = u(x, y) +iv(x, y) has an associated vector field F(x, y) =u(x, y)i+v(x, y)j.From this point of view, both F(x, y) = P(x, y)i+Q(x, y)jandf(z) =u(x, y) +iv(x, y) can be called vector fields.
This interpretation is often used to study various applications of complex functions in applied mathematical problems.
Letfbe a function defined in some neighborhood ofz0,with the possible exception of the pointz0itself. We say that thelimitoff(z) aszapproaches z0(independent of the path) is the numberw0if|f(z)−w0| →0 as|z−z0| → 0 and we write limz→z0f(z) =w0.Hence,f(z) can be made arbitrarily close tow0if we choosez sufficiently close toz0.Equivalently, we say thatw0 is the limit off aszapproachesz0if, for any given >0,there exists aδ >0 such that
0 < |z−z0| < δ =⇒ |f(z)−w0| < .
Figure 5.1 x
y
z
·
0·
0
z
u v
0
f(z) w0
·
Example 5.3.
By definition, we shall show that (i) limz→1−i(z2−2) =−2 + 2iand (ii) limz→1−i|z2−2|=√ 8.
(i). Given any >0,we have
|z2−2−(−2 + 2i)| = |z2−2i| = |z2+ 2i| = |z2+ 2i|
= |z−(1−i)||z+ (1−i)|
≤ |z−(1−i)|(|z−(1−i)|+ 2|1−i|)
≤ |z−(1−i)|(1 + 2√
2) if |z−(1−i)|<1
< if |z−(1−i)|<min
1,
1 + 2√ 2
. (ii). Given any >0,from (i) we have
||z2−2| −√
8| = ||z2−2| − | −2 + 2i||
≤ |z2−2−(−2 + 2i)|
< if |z−(1−i)|<min
1,
1 + 2√ 2
.
Example 5.4.
(i). Clearly, limz→z0z=z0. (ii). From the inequalities|Re(z−z0)| ≤ [(Re(z−z0))2+ (Im(z−z0))2]1/2 = |z−z0|,
|Im(z−z0)| ≤ |z−z0|,
it follows that limz→z0Rez= Rez0, limz→z0Imz= Imz0.
Example 5.5.
limz→0(z/z) does not exist. Indeed, we have limz→0
alongx-axis
z
z = lim
x→0
x+i0 x−i0 = 1,
lim z→0
alongy-axis
z
z = lim
y→0
0 +iy
0−iy = −1.
The following result relates real limits of u(x, y) and v(x, y) with the complex limit off(z) =u(x, y) +iv(x, y).
Theorem 5.1.
Letf(z) =u(x, y) +iv(x, y), z0 =x0+iy0, andw0 = u0+iv0.Then, limz→z0f(z) =w0if and only if limx→x0, y→y0u(x, y) =u0and limx→x0, y→y0v(x, y) =v0.
In view of Theorem 5.1 and the standard results in calculus, the follow- ing theorem is immediate.
Theorem 5.2.
If limz→z0f(z) = A and limz→z0g(z) = B, then (i) limz→z0(f(z)±g(z)) = A±B, (ii) limz→z0f(z)g(z) = AB, and (iii) limz→z0f(z) g(z) = A
B ifB= 0.
Complex Functions 31 For thecomposition of two functions f and g denoted and defined as (f◦g)(z) =f(g(z)),we have the following result.
Theorem 5.3.
If limz→z0g(z) =w0 and limw→w0f(w) =A,thenzlim→z0f(g(z)) = A = f
zlim→z0g(z)
.
Now we shall define limits that involve∞.For this, we note thatz→ ∞ means|z| → ∞,and similarly,f(z)→ ∞means|f(z)| → ∞.
The statement limz→z0f(z) =∞means that for any M >0 there is a δ >0 such that 0 <|z−z0| < δ implies|f(z)| > M and is equivalent to limz→z01/f(z) = 0.
The statement limz→∞f(z) = w0 means that for any > 0 there is an R >0 such that|z| > R implies |f(z)−w0|< , and is equivalent to limz→0f(1/z) =w0.
The statement limz→∞f(z) =∞means that for anyM >0 there is an R >0 such that|z|> Rimplies|f(z)|> M.
Example 5.6.
Since2z+ 3
3z+ 2 = 2 + 3/z 3 + 2/z,
limz→∞(2z+ 3)/(3z+ 2) = 2/3.Similarly, limz→∞(2z+ 3)/(3z2+ 2) = 0 and limz→∞(2z2+ 3)/(3z+ 2) =∞.
Letf be a function defined in a neighborhood ofz0.Then,f is contin- uousatz0if limz→z0f(z) =f(z0).Equivalently,f iscontinuousatz0if for any given >0,there exists aδ >0 such that
|z−z0| < δ =⇒ |f(z)−f(z0)| < .
A functionf is said to be continuous on a setS if it is continuous at each point ofS.
Example 5.7.
The functions f(z) = Re (z) and g(z) = Im (z) are continuous for allz.Example 5.8.
The functionf(z) =|z|is continuous for allz.For this, letz0be given. Thenzlim→z0|z| = lim
z→z0
(Rez)2+ (Imz)2 =
(Rez0)2+ (Imz0)2 = |z0|. Hence,f(z) is continuous atz0.Sincez0is arbitrary, we conclude thatf(z) is continuous for allz.
It follows from Theorem 5.1 that a functionf(z) = u(x, y) +iv(x, y) of a complex variable is continuous at a pointz0 =x0+iy0 if and only if u(x, y) andv(x, y) are continuous at (x0, y0).
Example 5.9.
The exponential function f(z) = ez is continuous on the whole complex plane sinceexcosyandexsiny both are continuous for all (x, y)∈IR2.The following result is an immediate consequence of Theorem 5.2.
Theorem 5.4.
If f(z) and g(z) are continuous at z0, then so are (i)f(z)±g(z),(ii)f(z)g(z),and (iii) f(z)/g(z) providedg(z0)= 0.Now letf :S→W, S1 ⊂S,andW1⊂W.Theinverse image denoted as f−1(W1) consists of all z ∈ S such that f(z) ∈ W1. It follows that f(f−1(W1))⊂W1 andf−1(f(S1))⊃S1.By definition, in terms of inverse image continuous functions can be characterized as follows: A function is continuous if and only if the inverse image of every open set is open.
Similarly, a function is continuous if and only if the inverse image of every closed set is closed.
For continuous functions we also have the following result.
Theorem 5.5.
Letf :S→Cbe continuous. Then,(i). a compact set ofS is mapped onto a compact set inf(S),and (ii). a connected set ofS is mapped onto a connected set off(S).
It is easy to see that the constant function and the function f(z) =z are continuous on the whole plane. Thus, from Theorem 5.4, we deduce that thepolynomial functions; i.e., functions of the form
P(z) = a0+a1z+a2z2+· · ·+anzn, (5.1) whereai, 0≤i≤nare constants, are also continuous on the whole plane.
Rational functionsin z,which are defined as quotients of polynomials; i.e., P(z)
Q(z) = a0+a1z+· · ·+anzn
b0+b1z+· · ·+bmzm, (5.2) are therefore continuous at each point where the denominator does not vanish.
Example 5.10.
We shall find the limits asz→2iof the functions f1(z) = z2−2z+ 1, f2(z) = z+ 2iz , f3(z) = z2+ 4 z(z−2i). Since f1(z) and f2(z) are continuous at z = 2i, we have limz→2if1(z) = f1(2i) =−3−4i,limz→2if2(z) =f2(2i) = 2.Since f3(z) is not defined at
Complex Functions 33 z= 2i,it is not continuous. However, for z= 2iandz= 0,we have
f3(z) = (z+ 2i)(z−2i)
z(z−2i) = z+ 2i
z = f2(z)
and so limz→2if3(z) = limz→2if2(z) = 2. Thus, the discontinuity off3(z) atz= 2ican be removed by settingf2(2i) = 2.The function f3(z) is said to have aremovable discontinuityatz= 2i.
Problems
5.1. For each of the following functions, describe the domain of defini- tion that is understood:
(a).f(z) = z
z2+ 3, (b).f(z) = z
z+z, (c). f(z) = 1 1− |z|2.
5.2. (a). Write the function f(z) = z3+ 2z+ 1 in the formf(z) = u(x, y) +iv(x, y).
(b). Suppose thatf(z) =x2−y2−2y+i(2x−2xy).Expressf(z) in terms ofz.
5.3. Show that when a limit of a function f(z) exists at a point z0,it is unique.
5.4. Use the definition of limit to prove that:
(a). lim
z→z0(z2+ 5) =z02+ 5, (b). lim
z→1−iz2= (1 +i)2, (c). lim
z→z0z=z0, (d). lim
z→2−i(2z+ 1) = 5−2i.
5.5. Find each of the following limits:
(a). lim
z→2+3i(z−5i)2, (b). lim
z→2
z2+ 3
iz , (c). lim
z→3i
z2+ 9 z−3i, (d). lim
z→i
z2+ 1
z4−1, (e). lim
z→∞
z2+ 1
z2+z+ 1−i, (f). lim
z→∞
z3+ 3iz2+ 7 z2−i . 5.6. Prove that:
(a). lim
z→0
z z
2
does not exist, (b). lim
z→0
z2 z = 0.
5.7. Show that if lim
z→z0f(z) = 0 and there exists a positive num- ber M such that |g(z)| ≤ M for all z in some neighborhood of z0, then
zlim→z0f(z)g(z) = 0.Use this result to show that limz→0zei/|z|= 0.
5.8. Show that if lim
z→z0f(z) =w0,then lim
z→z0|f(z)|=|w0|.
5.9. Suppose thatf is continuous atz0 and g is continuous at w0 = f(z0). Prove that the composite functiong◦f is continuous atz0.
5.10.Discuss the continuity of the function f(z) =
⎧⎨
⎩ z3−1
z−1 , |z| = 1 3, |z|= 1 at the points 1, −1, i,and−i.
5.11.Prove that the function f(z) = Arg(z) is discontinuous at each point on the nonpositive real axis.
5.12 (Cauchy’s Criterion).Show that limz→z0f(z) =w0if and only if for a given > 0 there exists a δ > 0 such that for any z, z satisfying
|z−z0|< δ, |z−z0|< δ,the inequality|f(z)−f(z)|< holds.
5.13.Prove Theorem 5.5.
5.14.The functionf :S→Cis said to beuniformly continuousonSif for every given >0 there exists aδ=δ()>0 such that|f(z1)−f(z2)|<
for all z1, z2 ∈ S with |z1−z2| < δ. Show that on a compact set every continuous function is uniformly continuous.
Answers or Hints
5.1. (a). z2 =−3 ⇐⇒ z =±√
3i, (b). z+z= 0 ⇐⇒ z is not purely imaginary; i.e., Re(z)= 0,(c). |z|2= 1 ⇐⇒ |z| = 1.
5.2. (a). (x+iy)3+ 2(x+iy) + 1 = (x3−3xy2+ 2x+ 1) +i(3x2y−y3+ 2y).
(b). Usex= (z+z)/2, y= (z−z)/2ito obtainf(z) =z2+ 2iz.
5.3. Suppose that limz→z0f(z) = w0 and limz→z0f(z) = w1. Then, for any positive number , there are positive numbers δ0 and δ1 such that
|f(z)−w0|< whenever 0<|z−z0| < δ0 and |f(z)−w1|< whenever 0 <|z−z0|< δ1.So, if 0 <|z−z0|< δ= min{δ0, δ1}, then|w0−w1|=
| −(f(z)−w0) + (f(z)−w1)| ≤ |f(z)−w0|+|f(z)−w1| < 2; i.e.,
|w0−w1|<2.But,can be chosen arbitrarily small. Hence,w0−w1= 0, or w0=w1.
5.4. (a).|z2+ 5−(z02+ 5)|=|z−z0||z+z0| ≤ |z−z0|(|z−z0|+ 2|z0|)
≤(1 + 2|z0|)|z−z0| if |z−z0|<1
< if 0<|z−z0|<min
1+2|z0|,1
,
(b).|z2−(1 +i)2|=|z2−(1−i)2| ≤ |z−(1−i)||z+ (1−i)|
≤ |z−(1−i)|(|z−(1−i)|+2|1−i|)<5|z−(1−i)| if |z−(1−i)|<1
Complex Functions 35
< if |z−(1−i)|<min{1, /5}, (c).|z−z0|=|z−z0|< if|z−z0|< ,
(d).|2z+1−(5−2i)|=|2z−(4−2i)|= 2|z−(2−i)|< if|z−(2−i)|< /2.
5.5. (a).−8i,(b).−7i/2,(c). 6i,(d).−1/2,(e). 1,(f). ∞. 5.6. (a). limz→0,z=x(z/z)2= 1, limz→0,y=x(z/z)2=−1.
(b). Let > 0. Choose δ=. Then, 0<|z−0|< δ implies |z2/z−0|=
|z|< .
5.7. Since limz→z0f(z) = 0,given any >0,there existsδ >0 such that
|f(z)−0|< /Mwhenever|z−z0|< δ.Thus,|f(z)g(z)−0|=|f(z)||g(z)| ≤ M|f(z)|< if|z−z0|< δ.
5.8. Use the fact||f(z)| − |w0|| ≤ |f(z)−w0|.
5.9. Let >0.Sinceg is continuous atw0,there exists aδ1>0 such that
|w−w0|< δ1 implies that |g(w)−g(w0)|< .Now, f is continuous atz0, so there exists aδ2>0 such that|z−z0|< δ2 implies|f(z)−f(z0)|< δ1. Combining these, we find that |z−z0| < δ2 implies |f(z)−f(z0)| < δ1, which in turn implies|(g◦f)(z)−(g◦f)(z0)|=|g[f(z)]−g[f(z0)]|< . 5.10.Continuous at 1,discontinuous at−1, i,−i.
5.11. f is not continuous at z0 if there exists 0 > 0 with the following property: For every δ > 0, there exists zδ such that |zδ −z0| < δ and
|f(zδ)−f(z0)| ≥0. Now letz0=x0<0. Take0= 3π/2.For eachδ >0, let zδ = x0−i(δ/2). Then, |zδ −z0| = |iδ/2| =δ/2 < δ, −π < f(zδ) <
−π/2, f(z0) =π,so |f(zδ)−f(z0)|>3π/2 =0, andf is not continuous atz0.Thus,f is not continuous at every point on the negative real axis. It is also not continuous atz= 0 because it is not defined there.
5.12.If f is continuous at z0, then given > 0 there exists a δ >0 such that |z1−z0| < δ/2 ⇒ |f(z1)−f(z0)| < /2 and |z2−z0| < δ/2 ⇒
|f(z2)−f(z0)| < /2. But then |z1−z2| ≤ |z1−z0|+|z0−z2| < δ ⇒
|f(z1)−f(z2)| ≤ |f(z1)−f(z0)|+|f(z2)−f(z0)|< .For the converse, we assume that 0<|z−z0|< δ, 0<|z−z0|< δ; otherwise, we can takez =z0
and then there is nothing to prove. Letzn →z0, zn =z0,and >0.There is aδ >0 such that 0<|z−z0|< δ, |z−z0|< δimplies|f(z)−f(z)|< , and there is an N such that n ≥N implies 0 <|zn−z0| < δ. Then, for m, n≥N,we have|f(zm)−f(zn)|< .So, w0= limn→∞f(zn) exists. To see that limz→z0f(z) =w0,take aδ1>0 such that 0<|z−z0|< δ1, 0<
|z−z0| < δ1 implies |f(z)−f(z)| < /2,and an N1 such that n ≥N1
implies 0<|zn−z0|< δ1 and|f(zn)−w0|< /2.Then, 0<|z−z0|< δ1
implies|f(z)−w0| ≤ |f(z)−f(zN1)|+|f(zN1)−w0|< .
5.13.(i). Suppose thatf :U →C is continuous andU is compact. Con- sider a covering off(U) to be open sets V.The inverse images f−1(V) are open and form a covering of U. Since U is compact, by Theorem 4.4 we can select a finite subcovering such that U ⊂ f−1(V1)∪ · · · ∪f−1(Vn). It follows that f(U) ⊂V1∪ · · · ∪Vn, which in view of Theorem 4.4 implies that f(U) is compact. (ii). Suppose that f :U →C is continuous andU is connected. Iff(U) =A∪B where AandB are open and disjoint, then U =f−1(A)∪f−1(B), which is a union of disjoint and open sets. SinceU
is connected, either f−1(A) =∅orf−1(B) =∅,and hence eitherA=∅ or B=∅.This implies thatf(U) is connected.
5.14.Use Theorem 4.4.