Professional Documents
Culture Documents
Analysis II - 5cor - Feb - 2011 (Ams) PDF
Analysis II - 5cor - Feb - 2011 (Ams) PDF
INTEGRAL CALCULUS
Preface 7
Chapter 1. Indefinite integrals (Primitives, Antiderivatives) 1
1. Definitions, some properties and basic formulas 1
2. Some results on polynomials 10
3. Primitives of rational functions 21
4. Primitives of irrational and trigonometric functions 27
5. Problems and exercises 47
Chapter 2. Definite integrals 49
1. The mass of a linear bar 49
2. Darboux sums and their applications 52
3. Lebesgue criterion and its applications 59
4. Mean theorem. Newton-Leibniz formula 63
5. The measure of a figure in Rn 68
6. Areas of plane figures bounded by graphics of functions 75
7. The volume of a rotational solid 85
8. The length of a curve in R3 87
9. Approximate computation of definite integrals. 90
10. Problems and exercises 100
Chapter 3. Improper (generalized) integrals 105
1. More on limits of functions of one variable 105
2. Improper integrals of the first type 112
3. Improper integrals of the second type 123
4. Problems and exercises 135
Chapter 4. Integrals with parameters 137
1. Proper integrals with parameters 137
2. Improper integrals with parameters 147
3. Euler’s functions gamma and beta 167
4. Problems and exercises 179
Chapter 5. Line integrals 181
1. The mass of a wire. Line integrals of the first type. 181
2. Line integrals of the second type. 194
3
4 CONTENTS
7
8 PREFACE
is why the reader must be grateful to her for the easy (but not short)
way I succeeded to present the improper integrals with parameters.
Finally, I mention the special help that trees and birds of my close
park gave me during my silent walk around the lake in those hot
evenings of August 2010. All of these are of a great importance when
one is thinking of mathematical affairs.
Prof. Dr. Sever Angel Popescu
Technical University of Civil Engineering Bucharest
Department of Mathematics and Computer Science
B=dul Lacul Tei 124, Sector 2, Bucharest, ROMANIA
angel.popescu@gmail.com
February, 2011
CHAPTER 1
Figure 1.
Such a figure is usually called a curvilinear trapezoid. The great
English scientist Sir Isaac Newton (1642-1727) discovered the mathe-
matical relation between the function y = f (x) and this area.
Theorem 1. (I. Newton) Let f : [a, b] → R+ be a continuous
function defined on a real closed interval [a, b] with nonnegative real
values. Let x be a number in (a, b] and let F (x) be the area of the
curvilinear trapezoid [AxMD] (the area under the graphic of f "up to
the point x" (see Fig.2). Then this area function is differentiable and
its derivative F 0 (x) at the point x is exactly f (x), the value of the initial
function f at the same point x.
1
2 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
Figure 2.
or
F (x) − F (x0 )
(1.2) f (x0 ) ≤ ≤ f (x).
(x − x0 )
Since f is continuous, taking limits when x → x0 in (1.2), we get that
F 0 (x0 ) exists and that it is equal to f (x0 ). ¤
The area function constructed above is called the Newton area func-
tion of f.
Corollary 1. Let f : [a, c) → R be a continuous function which
does not change its sign on the interval [a, c) (here c may be even ∞!).
Then, there is a differentiable function F defined on this interval (a, c)
such that F 0 (x) = f (x) for any x ∈ (a, c).
Proof. It is sufficient to take f to be nonnegative (otherwise re-
place f by −f ). Then we define like above F (x) = the area of the
curvilinear trapezoid [AxMD] (see Fig.2). ¤
Exercise 1. Try to substitute the above interval [a, c) for an open
finite or infinite interval (a, c). Moreover, try to substitute this last
interval for any open subset of R.
All of the discussion above is a real motivation for the following
general definition.
Definition 1. Let A be an open subset of the real line R. Let f :
A → R be a function defined on A with values in R. Any differentiable
function F : A → R such that its derivative F 0 (x) is equal to f (x) for
any x in A (simply if F 0 = f ) is called a primitive of f on A. It is also
called an antiderivative or an integral of f on A.
In the following all functions are defined only on subsets A of R,
which have no isolated points. We shall simply note A for a subset
satisfying this last property. We introduced this restriction because
it is a nonsense to speak about the limit of a function at an isolated
point. In particular, about its derivative!
If we remember the definition of the notion of a differential dF (a)
of a function F at a point a of A, namely dF (a) which is the linear
mapping defined on R with values in the same R, such that dF (a)(h) =
F 0 (a)h for any h in R, or dF (a) = F 0 (a)dx, where dx is the differential
of the variable x (in our case the identity mapping-see Analysis I, [Po]),
we can easily prove the following very useful result.
Theorem 2. Let f : A → R be a function defined on A (as above).
A function F : A → R is a primitive of f on A if and only if it is
4 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
for a point c in (a, b). Since H 0 (x) = 0 for any x of I, one gets that
H(b) = H(a). Let us denote by C this common value of H(x). Hence,
G − F = C, or G = F + C. This means that for this fixed number C,
we have that G(x) = F (x) + C for any x ∈ I. ¤
Remark 1. If the set A is not an interval, then the above result is
not always true. For instance, let A = (0, 1) ∪ (2, 3) and f (x) = 2x for
any x of A. It is easy to see that the following two functions
½
x2 , for x ∈ (0, 1)
F (x) = 2 ,
x + 1, for x ∈ (2, 3)
½ 2
x , for x ∈ (0, 1)
G(x) =
x2 , for x ∈ (2, 3)
are primitives of f. It is clear enough that there is no constant number
C such that G(x) = F (x) + C for any x in A. Indeed, for x ∈ (0, 1),
C = 0 and for x ∈ (2, 3), C = −1, so we cannot have the same constant
C on the entire set A.
Example 1. Let us find a primitive for the following continuous
function ⎧
⎨ x, if x ∈ [−1, 1]
f (x) = x3 , if x ∈ (1, 2) ,
⎩ x2 + 4, if x ∈ [2, 3)
³ 2 ´0 ³ 4 ´0 ³ 3 ´0
f : [−1, 3) → R. Since x2 = x, x4 = x3 and x3 + 4x = x2 +4,
any primitive of f is of the following form:
⎧ x2
⎨ 2 + C1 , if x ∈ [−1, 1]
x4
F (x) = + C2 , if x ∈ (1, 2) ,
⎩ x3 4
3
+ 4x + C3 , if x ∈ [2, 3)
Let us force now F to be continuous at x = 1 (F is differentiable, thus
it must be continuous):
1 1
+ C1 = + C2 ,
2 4
1
so we get that C2 = 4 + C1 . The continuity of F at x = 2 implies that
24 1 23
+ + C1 = + 8 + C3 ,
4 4 3
thus C3 = − 77
12
+ C1 . Hence, all the primitives of f are:
⎧ x2
⎨ 2
, if x ∈ [−1, 1]
x4
F (x) = + 14 , if x ∈ (1, 2) + C1 ,
⎩ x3 4 77
3
+ 4x − 12 , if x ∈ [2, 3)
6 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
Z
1
(1.4) dx = ln x, if x ∈ I,
x
where I is any interval contained in (0, ∞).
Z
1
(1.5) dx = ln(−x), if x ∈ J,
x
where J is any interval contained in (−∞, 0).
Z
(1.6) ex dx = ex
Z
ax
(1.7) ax dx = , if a > 0 and a 6= 1.
ln a
Z
1
(1.8a) dx = arctan x
x2 +1
Z
1 1 x
(1.9) dx = arctan , if a 6= 0.
x2 +a 2 a a
Z
(1.10) sin xdx = − cos x
Z
(1.11) cos xdx = sin x
Z
1
(1.12) √ dx = arcsin x, if x ∈ (−1, 1).
1 − x2
Z
1 x
(1.13) √ dx = arcsin , if x ∈ (−a, a).
a2 − x2 a
Z √
1
(1.14) √ dx = ln(x + x2 + α),
x2 + α
√
if α 6= 0, x2 + α > 0 and x + x2 + α > 0.
Z
1
(1.15) dx = tan x,
cos2 x
8 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
if x belongs to any interval which does not contain any number of the
form nπ + π2 , n ∈ Z.
Z
1
(1.16) dx = − cot x,
sin2 x
if x belongs to any interval which does not contain any number of the
form nπ, n ∈ Z.
Z Z
(1.17) sinh xdx = cosh x, cosh xdx = sinh x,
x −x x −x
where sinh x = e −e
2
and cosh x = e +e
2
.
All of the above formulas can be directly verified by differentiat-
ing the function from the right side of each equality. Let’s verify for
instance formula (1.14):
³ ´0 µ ¶
√ 1 2x 1
2
ln(x + x + α) = √ 1+ √ =√ .
x + x2 + α) 2 x2 + α x2 + α
Theorem 4. (linearity) Let f, g : I → R be two functionsR defined
on the
R same interval I and let α, β be two real numbers. Let f (x)dx
and
R g(x)dx beR two primitives of f and g respectively (on I). Then
α f (x)dx + β g(x)dx is a primitive of αf + βg on I.
¡R ¢0 ¡R ¢0
Proof. Since f (x)dx = f and g(x)dx = g (use only the
definition!), we get that
µ Z Z ¶0
α f (x)dx + β g(x)dx =
µZ ¶0 µZ ¶0
=α f (x)dx + β g(x)dx = αf + βg
and the proof is complete (why?). Here we used the linearity of the
differential operator h → h0 . ¤
R
This last result says that the "mapping" f → f (x)dx is linear.
Here is a concrete example of the way we can work with this last
property and with the above basic formulas.
R³ 5 √ 7
´
Example 4. Let us compute x − 3x + 2 3 x − √
5 3 dx. The
R x
linearity of the integral operator (see theorem 4) and formula (1.3)
imply that
Z µ ¶ Z Z
5
√
3
7 5
x − 3x + 2 x − √ 5
dx = x dx − 3 xdx+
x3
1. DEFINITIONS, SOME PROPERTIES AND BASIC FORMULAS 9
Z Z 4 2
1
− 35 x6 x2 x3 x5
+2 x dx − 7
3 x dx = −3 +2 4 −7 2 =
6 2 3 5
x6 3 2 3 4 35 2
= − x + x3 − x5 .
6 2 2 2
Let J be another interval and let u : J → I, x = u(t), be a function
of class C 1 (J). Then dx = u0 (t)dt (see Analysis I, [Po]). Let f : I → R
be a continuous function of the variable x ∈ I. Let F (x) be a primitive
of f on I. Then F (u(t)) is a primitive of the function f (u(t))u0 (t) of t,
i.e.
Z µZ ¶
(1.18) f (u(t))d(u(t)) = f (x)dx (u(t)).
Indeed,
[F (u(t))]0 = F 0 (u(t))u0 (t) = f (u(t))u0 (t).
Formula (1.18) is called the change of variable formula R for integral
computation. This formula says that if in an integral h(x)dx one can
put in evidence an expression
R u = u(x) such that h(x)dx = f (u(x))du
and if one can compute f (u)du then, put R instead of u, u(x) in this
last primitive and we obtain a primitive h(x)dx for the differential
form hdx.
R 1
Example 5. Let us compute 3x+2 dx, where 3x + 2 < 0. If we
denote
R 1 u = 3x + 2 then, du = 3dx (why?). Thus, our integral
R 1 becomes
1
3 3x+2
d(3x+2). Since u = 3x+2 < 0, a primitive for u du is ln(−u)
1
(see formula (1.5)). So, a primitive of 3x+2 dx is 13 ln(−3x − 2).
Sometimes it is easier to directly
R compute dx as a function of t and
dt. For instance, let us compute tan3 xdx. Let us change the variable:
t = tan x, x ∈ (− π2 , π2 ) (why this restriction?). Since
1 ¡ 2
¢
dt = dx = 1 + tan x dx = (1 + t2 )dx,
cos2 x
we get that
Z Z Z 3 Z Z
3 t3 t +t−t tdt
tan xdx = 2
dt = 2
dt = tdt − 2
=
1+t t +1 t +1
Z
t2 1 d(t2 + 1) t2 1
= − = − ln(t2 + 1) =
2 2 t2 + 1 2 2
1£ 2 ¤
= tan x − ln(tan2 x + 1) .
2
10 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
These last two formulas are called the integral by parts formulas.
R
When do we use them? If one has to compute the integral R h(x)dx
0 0
and if we can write h(x) = f (x)g (x) and if the integral f (x)g(x)dx
can be easier computed, then formula (1.19) works. For instance, if we
differentiate the function f (x) = ln x, we get x1 which is an "easier"
function from the point of view of integral calculus.
R Practically, let us
n
use this philosophy to compute the integral: In = x ln xdx, where n
is a natural number. For n = 0 we get
Z Z
I0 = ln xdx = (ln x)(x)0 dx.
and D is monic with the least degree such that D can be written as
in (2.11). In particular, if P and Q are coprime, then there are two
polynomials U0 and V0 such that
(2.12) 1 = P U0 + QV0
Proof. Let us define the following set of polynomials
(2.13)
S = {H = P U + QV : U and V are arbitrary polynomials in R[x]}.
It is easy to see that the sum between two polynomials of S is also a
polynomial in S.
If we multiply a polynomial of S by an arbitrary polynomial of R[x],
we also get a polynomial of S (prove these two last statements). Let
D be a (monic) nonzero polynomial of S of the least degree and let
M be another polynomial of S. Let us apply the Euclid’s algorithm for
dividing M by D.We get
M = CD + R,
where deg R < deg D. Since M, D ∈ S, then R = M − CD ∈ S. Thus,
R = 0 and so M = CD. Therefore D is a divisor of any polynomial of
S. In particular, D divides P and Q (P, Q ∈ S!) Since D ∈ S, there
are two polynomials U0 and V0 such that D = P U0 + QV0 .
=⇒) Let now G be the greatest common divisor of P and Q. Since
G divides P and Q, G also divides D. But D also divides P and Q,
thus the degree of D is at most equal to deg G. Since G divides D, we
have that deg G = deg D. Since D and G are both monic, we must
have that D = G. Thus, for the greatest common divisor we have the
relation (2.11).
⇐=) Let D0 be a monic nonzero polynomial of minimal degree such
that
D0 = P U1 + QV1 ∈ S.
Exactly like above we prove that D0 is the greatest common divisor of
P and Q.
The last statement is obvious. ¤
Practically, we can find the greatest common divisor and its ex-
pression (2.11) by transferring the problem from polynomials P and
Q to another pair of polynomials Q and R, where deg R < deg Q (if
deg P ≥ deg Q). Then to another pair R and R1 , where deg R1 <
deg R, etc. This idea directly comes from the Euclid’s division algo-
rithm. Indeed, assume that deg P ≥ deg Q (if deg Q > deg P, change P
with Q, etc.) and divide P by Q. There are polynomials C and R such
that P = CQ +R and deg R < deg Q. It is easy to see that the greatest
18 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
for k. Since P1n1 , P2n2 , ... , Pknk are monic irreducible and distinct one to
each other, P1n1 and L have no common roots. Indeed, if α is a root of
P1n1 and of L, then there is a j 6= 1 such that α is a root of P1 and of Pj ;
if α is real, P1 and Pj are both equal to x − α, a contradiction! (why?).
If α is not real, let α be the conjugate of α. Then P1 and Pj are both
equal to (x − α)(x − α), again a contradiction! (why?). Thus P1n1 and
L are coprime and we can apply theorem 9, formula (2.12). So there
are two uniquely defined polynomials V and U with deg V < deg L and
deg U < deg P1n1 such that
P1n1 V + LU = 1
(see theorem 10). Let us multiply both sides by Q P
= APQn1 L . We get:
1
∙ ¸
Q 1 QV QU
(2.19) = + n1 .
P A L P1
Let us put Q1 instead of QU and let us see that the number of irre-
ducible factors of L is k − 1. Applying the induction hypothesis we
obtain the representation (L is monic, so A = 1 in this case!):
QV Q2 Qk
(2.20) = n2 + ... + nk ,
L P2 Pk
where Q2 , Q3 , ..., Qk are uniquely defined polynomials. If we come back
to formula (2.19) with this expression of QV L
, we get exactly formula
(2.17). The other statement is a direct computational consequence of
this last formula. ¤
Example 6. Let us find the decomposition into simple fractions
of the rational function Q(x)
P (x)
= xx+1 2 2
4 +x2 . Since P (x) = x (x + 1), we
0 = A + D,
0 = B + C,
so A = 1, B = 1, C = −1 and D = −1. Therefore our decomposition
is
x+1 1 1 −x − 1
(2.22) 4 2
= 2+ + 2 .
x +x x x x +1
3. Primitives of rational functions
We shall use now the basic properties of the integrals (primitives)
and the above sketchy theory of rational functions in order to compute
their primitives.
Let us compute for instance a primitive R x+1 for the rational function
which
R appeared in example 6, namely x4 +x2 dx. We use the linearity
of in formula (2.22):
Z Z Z Z
x+1 1 1 −x − 1
dx = dx + dx + dx =
x4 + x2 x2 x x2 + 1
Z Z Z Z
−2 −1 1 2x 1
= x dx + x dx − 2
dx − 2
dx =
2 x +1 x +1
x−3 1
= + ln |x| − ln(x2 + 1) − arctan x,
−3 2
because
Z Z Z
2x d(x2 + 1) du
2
dx = 2
= = ln |u| = ln(x2 + 1).
x +1 x +1 u
Since any simple fraction has one of the following forms:
A
(3.1) , where A, α ∈ R,
x−α
A
(3.2) , where A, α ∈ R, n ∈ N∗ , n ≥ 2,
(x − α)n
Ax + B
(3.3) , where A, b, c ∈ R, ∆ = b2 − 4c < 0,
x2 + bx + c
Ax + B 2
(3.4) n , where A, b, c ∈ R, ∆ = b − 4c < 0, n ≥ 2,
(x2
+ bx + c)
we must show how to find a primitive for each of these cases.
Case 1
Z Z
A 1
(3.5) dx = A dx = A ln |x − α| .
x−α x−α
22 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
Case 2
Z Z
A 1
(3.6) dx = A dx =
(x − α)n (x − α)n
Z
(x − α)−n+1
A (x − α)−n dx = A , if n ≥ 2.
−n + 1
Case 3 Since
Z Z
Ax + B A 2x + b
2
dx = 2
dx+
x + bx + c 2 x + bx + c
µ ¶Z
Ab 1
+ − +B 2
dx
2 x + bx + c
and since Z
2x + b
2
dx = ln(x2 + bx + c),
x + bx + c
one can reduce everything to the computation of a primitive of the
following type:
Z Z
A 1
2
dx = A b 2
dx,
x + bx + c (x + 2 ) + a2
q
2
where a = c − b4 . Thus,
Z Z
A d(x + 2b )
(3.7) dx = A =
x2 + bx + c (x + 2b )2 + a2
1 x + 2b
= A arctan .
a a
R 1 1 x
Here we use the basic formula x2 +a 2 dx = a arctan a (see formula
(1.9)).
Case 4 Using a similar reasoning as in Case 3 one can reduce the
computation of the primitive of the simple fraction of formula (3.4) to
the following primitive:
Z Z
1 d(x + 2b )
dx = £ ¤n .
(x2 + bx + c)n (x + 2b )2 + a2
that
1 x
I1 = arctan .
a a
3. PRIMITIVES OF RATIONAL FUNCTIONS 23
Z Z
1 x2 1 1 x
− 2 n dx = In−1 − d(x2 + a2 ) =
a (x2 + a2 ) a2 2a2 (x2 + a2 )n
∙Z µ 2 ¶¸
1 1 (x + a2 )−n+1
= 2 In−1 − 2 xd .
a 2a −n + 1
Let us use now the formula of integrating by parts (see (1.20)) and
compute
Z µ 2 ¶ Z
(x + a2 )−n+1 (x2 + a2 )−n+1 1 1
xd =x − dx,
−n + 1 −n + 1 −n + 1 (x + a2 )n−1
2
Thus,
∙ ¸
1 1 (x2 + a2 )−n+1 1
In = 2 In−1 − 2 x − In−1 ,
a 2a −n + 1 −n + 1
or
∙ ¸
1 1 x
(3.8) In = 2 − 2 In−1 + 2 .
a 2a (n − 1) 2a (n − 1)(x2 + a2 )n−1
R dx
For instance, let us compute I2 = (x2 +a 2 )2 .
1 x x
(3.9) I2 = 3
tan−1 + 2 2 .
2a a 2a (x + a2 )
Formula (3.8) can be used to compute In "from up to down", i.e. the
computation of In reduces to the computation of In−1 . The computation
of In−1 reduces to the computation of In−2 , etc., up to the computation
of I1 = a1 arctan xa .
In the following examples we use the ideas and experience just ex-
posed in the above four cases.
3
Example 7. What is a primitive of 4x+5 , where x runs over an
5
interval I which does not contain x = − 4 .
Z Z Z
3 3 d(4x + 5) 3 du 3 3
dx = = = ln |u| = ln |4x + 5| .
4x + 5 4 4x + 5 4 u 4 4
1
Example 8. Let us compute a primitive of f (x) = (2x+5) 5 , where
5
x belongs to an interval I which does not contain x = − 2 .
Z Z Z
1 1 −5 1
dx = (2x + 5) d(2x + 5) = u−5 du =
(2x + 5)5 2 2
24 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
Z Z
x+3 1 2x + 4
I= 3 dx = dx+
2
(x + 4x + 6) 2 (x + 4x + 6)3
2
Z Z 0
1 1 (x2 + 4x + 6)
(3.10) + dx = dx+
(x2 + 4x + 6)3 2 (x2 + 4x + 6)3
Z Z Z
1 1 −3 dv
+ 3 d(x + 2) = u du + £ √ ¤3 ,
[(x + 2)2 + 2] 2 v2 + ( 2)2
where u = x2 + 4x + 6 and v = x + 2. Let
Z
dv
I3 = £ √ ¤3
v 2 + ( 2)2
3. PRIMITIVES OF RATIONAL FUNCTIONS 25
Thus,
Z Z Z
1 1
(3.11) dx = b−2 −2
x dx − b −2
dx =
x (x + b2 )
2 2 x2 + b2
x−1 1 x
= b−2 − b−2 arctan .
−1 b b
R
Case 3. a 6= 0, b = 0. Then I = x2 (x21+a2 ) dx. Use now formula
(3.11) with a instead of b and find
Z −1
1 −2 x −2 1 x
dx = a − a arctan .
x2 (x2 + a2 ) −1 a a
1
Case 4. a 6= 0, b 6= 0, a 6= b. Let us take the fraction (x2 +a2 )(x2 +b2 ) and
Finally,
Z
1
dx =
(x + a )(x2 + b2 )
2 2
∙Z Z ¸
1 1 1
= 2 dx − dx =
b − a2 x + a2
2 x2 + b2
∙ ¸
1 1 x 1 x
= 2 arctan − arctan .
b − a2 a a b b
R 1
Case 5. a = b 6= 0. Then I = (x2 +a 2 )2 = I2 , with the notation of
X
n X
m
P (x, y) = aij xi y j = a00 + a10 x + a01 y + ... + anm xn y m ,
i=0 j=0
(4.10) ¯
¯ √ ¯¯ ¯
¯ √ ¯¯
= −2A ln ¯t + 1 + 2¯ − 2B ln ¯t + 1 − 2¯ − C ln(t2 + 1) − 2D arctan t,
where A, B, C, D have the numerical values from formula (4.9). From
(4.8) we get √
4 − x2 − 2
t= .
x
With this last value of t we come in (4.10) and obtain
¯√ ¯
¯ 4 − x2 − 2 √ ¯
I = −2A ln ¯¯ + 1 + 2¯¯ −
x
¯√ ¯
¯ 4 − x2 − 2 √ ¯
−2B ln ¯¯ + 1 − 2¯¯ −
x
Ã∙ √ ¸2 ! √
4 − x2 − 2 4 − x2 − 2
−C ln + 1 − 2D arctan .
x x
Remark 2. To compute the primitive B1 we also p can use trigono-
R
metric substitutions. For instance, if in I = R(x, x2 + δ 2 )dx we
put x = δ tan t, we get
Z
δ δ 1
dx = 2
dt, I = R(δ tan t, )δ 2 dt.
cos t |cos t| cos t
This last primitive is a rational function of sin x and cos x. We shell
see later how to integrate a rational function of sin x and cos x.
Example 15. Let us make x = a tan t (a > 0) and compute
Z √ Z Z Z
2 2 2 1 2 d(sin t) 2 du
I= x + a dx = a dt = a =a ,
3
cos t 4
cos t (1 − u2 )2
where u = sin t. Since
1 A B C D
2 = 2 + + 2 + ,
(1 − u2 ) (1 − u) 1 − u (1 + u) 1+u
we find A = B = C = D = 14 . So
∙Z Z Z Z ¸
a2 du du du du
I= + + + =
4 (1 − u)2 1−u (1 + u)2 1+u
∙ ¸ µ ¶
a2 1 1 a2 2u 1+u
− ln(1 − u) − + ln(1 + u) = + ln .
4 1−u 1+u 4 1 − u2 1−u
Coming back to t we obtain
µ ¶
a2 2 sin t 1 + sin t
I= + ln .
4 cos2 t 1 − sin t
32 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
Proof. We prove only the fact that if we are in one of the three
cases above, then the integral can be "rationalized", i.e. there is a
sequence of "rational substitutions" of variables such that our binomial
differential form xm (axn +b)p dx becomes a rational differential form, i.e.
a differential form of the type Q(u)du, where Q is a rational function in
a new variable u. The reverse part of the theorem cannot be proved by
elementary tools. It involves deep knowledge of Algebraic Geometry.
First of all let us make the natural substitution: xn = t (if n 6= 1; if
n = 1 we pass directly to the second step of the following reasoning).
1 1
Thus x = t n and dx = n1 t n −1 dt. Hence, our binomial integral becomes
Z
1
(4.12) tq (at + b)p dt,
n
u
which is a primitive of a rational function. We decompose (u3 −4) 4 into
If q = m+1
n
− 1 is not an integer, i.e. if m+1 n
is not an integer, q =
q1
q2
, simplified and q1 , q2 integers with q2 6= 1, then the obvious substitu-
tion t = uq2 makes the binomial differential form tq (at + b)p dt rational.
Indeed,
tq (at + b)p dt = uq1 (auq2 + b)p q2 uq2 −1 du
is rational because q1 , q2 and p are integers.
Case 2. p is not an integer, but m+1 n
is an integer. Let p = pp12 ,
simplified and p1 , p2 integers, with p2 6= 1. Then the substitution
at + b = up2 give rise to a rational differential form. Indeed, this time
p p p
q = m+1 n
− 1 is an integer, t = u 2a−b and dt = a2 u 2 −1 du, so
µ p2 ¶q p2
q p u −b p
t (at + b) dt = up1 u 2 −1 du
a a
is rational because p1 , p2 and q are integers.
Case 3. p = pp12 is not an integer and m+1 n
is not an integer too.
In this case we make the following "trick" in the canonical differ-
ential form:
µ ¶p
q p q+p at + b
(4.13) t (at + b) dt = t dt.
t
We apply now the same idea like above.
If q + p = m+1
n
− 1 + p is an integer, i.e. if m+1 n
+ p is an integer,
at+b p2
then the obvious substitution t = u , where p2 is the denominator
of p, leads to a rational differential form. Indeed, this time
b
t = p2 = b (up2 − a)−1 ,
u −a
dt = −bp2 up2 −1 (up2 − a)−2 du,
so
µ ¶p
at + b
q+p
t dt = bq+p (up2 − a)−(q+p) up1 (−bp2 ) up2 −1 (up2 − a)−2 du.
t
Since q + p, p1 and p2 are integers, then this last differential form
is rational and the implication "⇐=" of the theorem is completely
proved. ¤
Remark 4. The case 3 which naturally appeared during the proof
of the above theorem 12 can also be manipulated in the following way.
Instead of the "trick" used in formula (4.13) we can use the following
one:
µ ¶q
q p t
(4.14) t (at + b) dt = (at + b)p+q dt.
at + b
4. PRIMITIVES OF IRRATIONAL AND TRIGONOMETRIC FUNCTIONS 35
Z
1 £ ¤0
− u5 (2 + 3u3 )−2 du =
3
½ Z ¾
1 5 3 −2 4 3 −2
− u (2 + 3u ) − 5 u (2 + 3u ) du =
3
½ Z ¾
1 5 3 −2 5 2
£ ¤
3 −1 0 by parts
− u (2 + 3u ) + u (2 + 3u ) du =
3 9
½ ∙ Z ¸¾
1 5 3 −2 5 2 3 −1 u
(4.15) − u (2 + 3u ) + u (2 + 3u ) − 2 du .
3 9 2 + 3u3
This last integral
Z Z Z
u 1 u 1 u
du = 2 du = du,
2 + 3u3 3 3
+u3 3 a3 + u3
36 1. INDEFINITE INTEGRALS (PRIMITIVES, ANTIDERIVATIVES)
q
where a = 3 23 , is an integral of a rational function. We can integrate it
u
by the usual methods of decomposing of the integrand a3 +u 3 into simple
fractions:
u 1 1 1 u+a
3 3
=− + .
a +u 3a u + a 3a u − ua + a2
2
Hence,
Z Z Z
1 u 1 1 1 2u − a + 3a
3 3
du = − du + du =
3 a +u 9a u+a 18a u2 − ua + a2
Z
1 1 ¯ 2 ¯ 1 1
(4.16) − ln |u + a| + ¯
ln u − ua + a + 2¯
du.
9a 18a 6 u − ua + a2
2
1 ¯¯ 1 ¯ ¯ 1 ¯ 1
2 ¯ 1 ¯ 1
2 ¯ 1 −1 2x 6 − a
− − ln ¯x 6 + a¯ + ln ¯x 3 − x 6 a + a ¯ + √ tan √ ]}.
3 9a 18a 3a 3 a 3
q
where a = 3 23 .
etc.
R √
Example 17. Let us compute the integral I = x3 3 1 + x2 dx.
Since m = 3, n = 2 and p = 13 , p is not an integer and m+1 n
= 2
is an integer, we are in Case 2. We can see that it is possible to put
directly
√
1 + x2 = u3 , u = 1 + x2
3
(4.20)
(without making first of all x2 = t, why?). Thus,
3u2
2xdx = 3u2 du =⇒ dx = du.
2x
√
Let us come back to our differential form x3 3 1 + x2 dx and perform
these new substitutions:
√ 3u2 3 3
x3 1 + x2 dx = x3 u du = x2 u3 du = (u3 − 1)u3 du.
3
2x 2 2
So
Z µ ¶
3 3 3 3 u7 u4 3 ¡ ¢7 3 ¡ ¢4
(u − 1)u du = − = 1 + x2 3 − 1 + x2 3 .
2 2 7 4 14 8
R 1+x2
√
Example 18. Let us compute now I = x4
dx. This is a bino-
1
mial integral with m = −4, n = 2 and p = 2 . Since p is not an integer,
m+1
n
= − 32 is also not an integer, but m+1
n
+ p = −1 is an integer (Case
2
3), we can directly make the substitution 1+x x2
= u2 (because t = x2 in
formula (4.13)). So
1
1 + x2 = u2 x2 , x2 = , 2xdx = 2x2 udu + 2u2 xdx,
u2 −1
x2 u xu
dx = 2
du = du.
x−u x 1 − u2
4. PRIMITIVES OF IRRATIONAL AND TRIGONOMETRIC FUNCTIONS 39
Thus, our differential form becomes (we consider only the case x > 0
and u ∈ (1, ∞)):
√
1 + x2 ux xu u2
dx = du = du =
x4 x4 1 − u2 x2 (1 − u2 )
u2
= ¡ 1 ¢ 2)
du = −u2 du.
u2 −1
(1 − u
Finally we get:
Z √ Z 3
µ ¶3
2 2
1 + x2 u 1 1 + x
dx = − u2 du = − = − .
x4 3 3 x2
The best idea is to make "formal" computations (not taking count of
the definition domains of different expressions which appear during sub-
stitutions!) and, in the end, to verify by a direct differentiation the
obtained result. In our example,
" µ ¶ 3 #0 √
1 1 + x2 2 1 + x2
− = , x 6= 0.
3 x2 x4
Hence, in spite of the required limitations which appeared during
√
a par-
2
ticular computation, a primitive of the differential form 1+x x4
dx, on
³ 2 ´ 32
each interval which does not contain 0, is − 13 1+x
x2
.
integral becomes: Z
1 3 1
t− 4 (1 + t) 2 dt.
4
The last integral is a binomial integral with m = − 34 , n = 1 and p = 12 .
Since p is not an integer, m+1n
= 14 is also not an integer and m+1
n
+p
= 34 is not an integer, we are in no one of the three cases of the Cebysev
theorem. So we cannot reduce this primitive to a rational primitive by
rational substitutions. O. K., but the question is still alive! √Is there
an elementary differentiable function F (x) such that F 0 (x) = 1 + x4 ?
The answer is no, but we need a lot of higher Mathematics to prove it!
R
D. Primitives of the form R(cos x, sin x)dx, where R(x1 , x2 )
is a rational function
The general method consists in the following change of variable.
Assume that x belongs to an interval J on which the function x → tan x2
is invertible and its inverse is of class C 1 . Since
1 − tan2 x2 2 tan x2
cos x = , sin x = ,
1 + tan2 x2 1 + tan2 x2
it is naturally to put t = tan x2 , x = 2 arctan t, dx = 1+t
2
2 dt, so
Z Z µ ¶
1 − t2 2t 2
R(cos x, sin x)dx = R , dt,
1 + t 1 + t 1 + t2
2 2
where u = cos x, or
Z Z
(4.22) R(cos x, sin x)dx = T (sin x) cos xdx =
Z Z
= T (sin x)d(sin x) = T (v)dv,
where v = sin x.
R
Example 20. If we want to compute I = sin4 x cos3 xdx by the
general substitution t = tan x2 , we get a very complicated integral of a
rational function in t (why is it so complicated?)
Z µ ¶4 µ ¶3
2t 1 − t2 2
I= 2 2
dt.
1+t 1+t 1 + t2
But, if we use the substitution described in (4.22), we get:
Z Z
I = sin x cos x cos xdx = sin4 x(1 − sin2 x)d(sin x) =
4 2
Z
v5 v7 sin5 x sin7 x
= v4 (1 − v 2 )dv = − = − .
5 7 5 7
Remark 8. Suppose now that cos x and sin x appear to even powers
in
R(cos x, sin x). Then always one can write R(cos x, sin x)dx as
1
S(tan x) dx = S(tan x)d(tan x) = S(u)du,
cos2 x
where u = tan x and S is a rational function of u (why all of these?).
The substitution
R u = tan x can also be done even in other cases. For
instance, sin3sin xdx
x+cos3 x
dx can be "rationalized" by puting tan x = u.
R cos2 x
Example 21. Let us compute I = 1+sin 2 x dx by this last method
x
(the general substitution t = tan 2 is not good at all!-why?). Since
1 + tan2 x = cos12 x , and since u = tan x implies that x = tan−1 u and
1
dx = 1+u 2 du, we get
Z Z Z
cos2 x 1 1
2 dx = 1 2
dx = du.
1 + sin x cos2 x
+ tan x (1 + 2u ) (1 + u2 )
2
1 1
= (cos x + i sin x) = exp(ix).
i i
We see from here that the usual rules for computing primitives of the
real valued functions extends naturally to complex valued functions of
real variables. This is true because such extension works in the case of
the differential calculus (see Analysis I, [Po]). So we can directly write
Z Z Z
1 1 1
exp(ix)dx = exp(ix)d(ix) = exp udu = exp u =
i i i
1
= exp(ix) = −i exp(ix).
i
4. PRIMITIVES OF IRRATIONAL AND TRIGONOMETRIC FUNCTIONS 43
and
exp(ax)
I(a, b) = [a sin(bx) − b cos(bx)] .
a2 + b2
Remark 9. A way to compute primitives of the form
Z
R(cos mx, sin nx)dx,
exp(inx) − exp(−inx)
(4.24) sin nx =
2i
and to reduce the computation to a primitive of a complex valued func-
tion (of a real variable).
Example 23. Let us compute a primitive for the trigonometric
differential form sin 3x cos4 xdx.
Since sin 3x = exp(3ix)−exp(−3ix)
2i
and since
1 + cos 2x 1 exp(2ix) + exp(−2ix)
cos2 x = = + =
2 2 4
1
= [2 + exp(2ix) + exp(−2ix)] ,
4
1
cos4 x = [6 + exp(4ix) + exp(−4ix) + 4 exp(2ix) + 4 exp(−2ix)] ,
16
one has
Z Z
4 1
sin 3x cos xdx = [exp(3ix) − exp(−3ix)] ×
32i
× [6 + exp(4ix) + exp(−4ix) + 4 exp(2ix) + 4 exp(−2ix)] dx =
Z
1
= [6 exp(3ix) + exp(7ix) + exp(−ix) + 4 exp(5ix) + +4 exp(ix)−
32i
−6 exp(−3ix) − exp(ix) − exp(−7ix) − 4 exp(−ix) − 4 exp(−5ix)]dx =
Z
1
= [6 sin(3x) + sin(7x) + 3 sin x + 4 sin 5x]dx =
16
1 1 3 1
− cos 3x − cos 7x − cos x − cos 5x.
8 112 16 20
4. PRIMITIVES OF IRRATIONAL AND TRIGONOMETRIC FUNCTIONS 45
R
Exercise 2. Let In = sinn xdx, where n > 2, and n = 2k is even
(k > 1). If n is odd it is easier to write
n−1
sinn xdx = − sinn−1 x d(cos x) = −(1 − u2 ) 2 du,
etc. Let us compute In by firstly finding a recurrence formula:
Z Z
2k
¡ 2k−2 ¢
I2k = sin xdx = sin x (1 − cos2 x)dx =
Z
by parts
I2k−2 − sin2k−2 x cos xd(sin x) =
So Z
sinn xdx =
Z µ ¶Z
1 n
= n n [ exp(inx)dx − exp i(n − 2)xdx+
2 i 1
µ ¶Z
n
+ exp i(n − 4)xdx − ...].
2
For instance,
Z ∙Z Z ¸
4 1
sin xdx = exp(4ix)dx − 4 exp(2ix)dx +
16
∙Z Z Z ¸
6
+ dx − 4 exp(−2ix)dx + exp(−4ix)dx =
16
" #
1 exp(4ix) exp(2ix) exp(−2ix) exp(−4ix)
4i −4 +6x − 4 + −4i =
16 | {z 2i } −2i
| {z }
∙ ¸
1 1
sin 4x − 4 sin 2x + 6x .
16 2
An elementary trigonometric computation tells us that this expression
is exactly that one from (4.26). Moreover, it is more beautiful, because
it has no powers of trigonometric functions! These last ones are "not
desirable" during the integration computation.
Remark 10. It is very useful to knowR that the Rfollowing primitives
R exp R
x
are not elementary functions: dx; sinx x dx; cosx x dx; sinh x
dx;
R cosh x R 2
R 2
R x
2
R 1
R dϕ
x
x
dx; sin x dx; cos x dx; exp(−x )dx; ln x dx; √ 2 2 ;
1−k sin ϕ
Rp 2
R dϕ
2
1 − k sin ϕdϕ; 2
√ 2
, where k, h ∈ (−1, 1) \
(1+h sin ϕ) 2
1−k sin ϕ
{0} (elliptic integrals). Proofs for such statements belong to very high
Mathematics. It implies deep knowledge of Algebraic Geometry and Al-
gebraic Functions Theory. We note that not all the above R statements
are "independent" one to eachRother". For instance, if expx x dx is not
1
an elementary function, thenR 1 ln x dx is also not an elementary func-
tion. Indeed, let F (x) = ln x dx, a primitive of the differential form
1
ln x
dx. Let us make the substitution x = exp(t) in this last differential
form:
1 exp t
dx = dt.
ln x t
So a primitive of expt t dt is F (exp(t)) on any interval I which does
not contain 0 (prove it!). But, if F were elementary, then F (exp(t)
5. PROBLEMS AND EXERCISES 47
Definite integrals
the variation of f over the interval [xi−1 , xi ], i.e. the length of the inter-
val f ([xi−1 , xi ]) if f is continuous. Recall that a function f : [a, b] → R
is uniformly continuous if for any small real number ε > 0 there is a real
number δ ε > 0 such that whenever x0 , x00 are in [a, b] and |x0 − x00 | < δ ε ,
one has that |f (x0 ) − f (x00 )| < ε. It is clear enough that a uniformly
continuous function is also a continuous one (why?). Since our con-
tinuous function f is also uniformly continuous on [a, b] (see [Po], Th.
59), for any small number ε > 0, one can find a δ ε > 0, such that for
any division
∆ : a = x0 < x1 < x2 < ... < xn = b
with k∆k < δ ε , the variation ω f [xi−1 , xi ] is less then ε. This means that
one can well approximate the density function f on the subinterval
[xi−1 , xi ] with any fixed value f (ξ i ) of f at a fixed point ξ i ∈ [xi−1 , xi ].
The set of the fixed points {ξ i }, ξ i ∈ [xi−1 , xi ], i = 1, 2, ..., n is called
a set of marking points for the division ∆. Thus, the mass of the bar
can be well approximated by the sum:
X
n
(1.2) S(f ; ∆; {ξ i }) = f (ξ i )(xi − xi−1 ).
i=1
Figure 1.
Thus, these sums can well approximate the area of the trapezoid
aABb (see Fig.1) If the set of all these Riemann sums (when ∆ and {ξ i }
vary such that k∆k → 0) have one and only one limit point I(f ), we say
that this number I(f ) is exactly the mass of the bar (think of this intu-
itively!). Here it was an example which justify the following definition.
We recall that a limit point for a nonempty subset A of R is a point
a of R such that for any small real number ε > 0 the ε-neighborhood
(a − ε, a + ε) contains an infinite number of elements which are in A.
For instance, {0} is a limit point of the set A = {1, 1/2, 1/3, ...1/n, ...}.
Definition 3. We say that a function f : [a, b] → R is (Riemann)
integrable on [a, b] if there exist a real number I(f ), called the definite
Rb
integral of f from a to b and denoted by a f (x)dx, such that for any
ε > 0 there is a δ ε > 0 with the following property: if ∆ is a division
of [a, b] with k∆k < δ ε , then |S(f ; ∆; {ξ i }) − I(f )| < ε for any set of
marking points {ξ i } of the division ∆.
Rb
The above defined number I(f ) = a f (x)dx can be well approxi-
mated with Riemann sums of type (1.2). It is easy to see that such a
number I(f ) is unique (do it!). The above definition says that if we
look at a Riemann sum as a function of the division ∆ and of {ξ i }, a
set of marking points of ∆, this last defined function "has a limit" as
k∆k → 0 if and only if f is Riemann integrable on [a, b].
1. THE MASS OF A LINEAR BAR 51
X
n X
n
S(f ; ∆; {ξ i }) = f (ξ i )(xi − xi−1 ) = k (xi − xi−1 ) = k(b − a),
i=1 i=1
thus all the Riemann sums are equal to the constant number k(b − a),
i.e. f is integrable on [a, b] and its integral I(f ) = k(b − a).
Even for not complicated functions, to prove the integrability and
to compute directly I(f ), usually is not an easy task.
Example 27. For instance, let f (x) = x, x ∈ [0, 1]. Then
X
n X
n
S(f ; ∆; {ξ i }) = f (ξ i )(xi − xi−1 ) = ξ i (xi − xi−1 ).
i=1 i=1
Since ξ i ∈ [xi−1 , xi ], i = 1, 2, ..., n, we can write
X n X n X
n
xi−1 (xi − xi−1 ) ≤ ξ i (xi − xi−1 ) ≤ xi (xi − xi−1 ).
i=1 i=1 i=1
Since " #
X
n
1 Xn
xi (xi − xi−1 ) = 1+ (xi − xi−1 )2
i=1
2 i=1
and " #
X
n
1 X n
xi−1 (xi − xi−1 ) = 1− (xi − xi−1 )2 ,
i=1
2 i=1
P
it is enough to see that ni=1 (xi − xi−1 )2 → 0, wheneverPn k∆k → 0.2 If
the division is an equidistant one (xi −xi−1 = 1/n), i=1 (xi −xi−1 ) =
1/n → 0, whenever k∆k = 1/n → 0, i.e. when n → ∞. For arbitrary
divisions of [0, 1] to see this is not so easy. Finally, by using the future
derived Leibniz-Newton formula (see theorem 23), one R 1 has that our
simple function f (x) = x, x ∈ [0, 1] is integrable and 0 xdx = 1/2.
The main problems we are concerned with are: 1) When a function
f : [a, b] → R is Riemann integrable? and 2) If f is integrable, how
Rb
do we compute its definite integral a f (x)dx? Let us begin with the
following remark.
Theorem 13. Let f : [a, b] → R be a Riemann integrable function.
Then it is bounded, i.e. kf k = sup |f (x)| is a finite number.
x∈[a,b]
52 2. DEFINITE INTEGRALS
Rb
Proof. Let I(f ) = a f (x)dx and let ε > 0 be a small number,
ε = 1/10 for instance. Let ∆ : a = x0 < x1 < x2 < ... < xn = b be a
division of [a, b] such that
(1.3) |I(f ) − S(f ; ∆; {ξ i })| < 1/10,
for any set of marking points {ξ i } of ∆. Suppose that f is unbounded
to ∞. Then f is unbounded at least on a subinterval [xj−1 , xj ]. Let
(1) (2) (k)
us change the marking point ξ j with ξ j , ξ j , ..., ξ j , ... such that
(k)
f (ξ j ) → ∞, when k → ∞. But
" n #
X (k)
S(f ; ∆; {ξ i }) = f (ξ i )(xi − xi−1 ) + f (ξ j )(xj − xj−1 ) → ∞,
i=1,i6=j
Figure 2.
Figure 3.
(see (2.2)), there is the greatest lower bound I ∗ (f ) of {S∆ }. This last
number is called the superior Darboux integral of f. There are examples
for which these two Darboux integrals are not equal. For instance, let
f : [a, b] → R be the mapping which associates to a rational number
of [a, b] the value 0 and to an irrational number of [a, b] the value 1.
Then, I∗ (f ) = 0 and I ∗ (f ) = 1. To make things clearer, let us consider
two divisions ∆ and ∆0 of [a, b] such that ∆0 contains the points of ∆
and maybe some additional other points. We write this as: ∆ ≺ ∆0 .
Adding successively to ∆ only one point c at a time, we can easily
54 2. DEFINITE INTEGRALS
Figure 4.
Figure 5
Taking now two arbitrary divisions ∆0 and ∆00 of [a, b], we can
construct the "least" division ∆000 = ∆0 ∪ ∆00 which contains ∆0 and
∆00 at the same time: ∆0 ≺ ∆000 and ∆00 ≺ ∆000 . Thus (2.3) becomes:
s∆0 ≤ s∆000 ≤ S∆000 ≤ S∆00 ,
so s∆0 ≤ S∆00 for any two arbitrary divisions ∆0 and ∆00 of [a, b].
Thus the inequality I∗ (f ) ≤ I ∗ (f ) is clear! Moreover, since mi =
inf f (x), keeping the division ∆ fixed, one can find a sequence
x∈[xi−1 ,xi ]
2. DARBOUX SUMS AND THEIR APPLICATIONS 55
(n) (n)
{ξ i } of numbers in [xi−1, xi ] such that f (ξ i ) → mi as n → ∞. Thus,
as close as we want to s∆ one can find Riemann sums of the form
(n)
S(f ; ∆; {ξ i }) for a fixed ∆. The same is also true for S∆ . Let us assume
Rb
now that f : [a, b] → R is Riemann integrable with I(f ) = a f (x)dx
and let ε > 0 be a small real number. Let δ ε > 0 be small enough such
that if k∆k < δ ε then
ε
(2.4) |I(f ) − S(f ; ∆; {ξ i })| <
4
for any marking points {ξ i } of the division ∆. Take now two sets of
marking points {ξ 0i } and {ξ 00i } with the property that:
ε ε
(2.5) |S∆ − S(f ; ∆; {ξ 0i })| < and |S(f ; ∆; {ξ 00i }) − s∆ | < .
4 4
So, one can write:
S∆ − s∆ ≤ |S∆ − S(f ; ∆; {ξ 0i })| + |S(f ; ∆; {ξ 0i }) − I(f )| +
ε ε ε ε
+ |I(f ) − S(f ; ∆; {ξ 00i })| + |S(f ; ∆; {ξ 00i }) − s∆ | < + + + = ε.
4 4 4 4
Thus S∆ − s∆ < ε for any division ∆ with k∆k < δ ε . We just proved
an implication of the following basic result.
Theorem 14. (Darboux Criterion) f : [a, b] → R is Riemann in-
tegrable if and only if I∗ (f ) = I ∗ (f ), i.e. if and only if for any small
real number ε > 0, there is another small real number δ ε > 0 such that
if ∆ is any division with k∆k < δ ε , one has that S∆ − s∆ < ε.
Proof. It remains to prove the converse. We denote the common
value I∗ (f ) = I ∗ (f ) by I(f ) and let us prove that I(f ) is the unique
limit point of all the Riemann sums S(f ; ∆; {ξ i }), when k∆k → 0. Let
again ε > 0 be a small real number and let δε > 0 be the corresponding
small real number such that if ∆ is any division with k∆k < δ ε , one
has that S∆ − s∆ < ε. Since s∆ ≤ S(f ; ∆; {ξ i }) ≤ S∆ (see 2.2) and
since s∆ ≤ I∗ (f ) = I ∗ (f ) = I(f ) ≤ S∆ one obtains that
|S(f ; ∆; {ξ i }) − I(f )| ≤ S∆ − s∆ < ε
for any division ∆ with k∆k < δ ε and for any set of marking points
(ξ i ) of ∆. ¤
Remark 11. (Riemann Criterion) It is not difficult to prove that
f : [a, b] → R is Riemann integrable if and only if there is a real
number I(f ) such that for any sequence {∆n } of divisions of [a, b] with
(n)
k∆n k → 0 as n → ∞ and for any set of marking points {ξ i } of ∆n ,
(n)
n = 1, 2, ..., one has that S(f ; ∆n ; {ξ i }) → I(f ), when n → ∞. Thus,
if we know that f is integrable on [a, b], we can use some special types
56 2. DEFINITE INTEGRALS
ε
< (x1 − a + x2 − x1 + x3 − x2 + ... + b) = ε.
b−a
i.e. f is Riemann integrable on [a, b]. ¤
a
f (x)dx exists and
Z b Z c Z b
(2.10) f (x)dx = fe(x)dx + fe(x)dx.
a a c
58 2. DEFINITE INTEGRALS
For instance,
½ 1
x, x 6= 2
g(x) = 1 , g : [0, 1] → R,
0, x = 2
is integrable and
Z 1 Z 1
1
g(x)dx =
xdx = ,
0 0 2
as we just proved in example 27. The morale is that if the function is
bounded, we do not care with it at a finite number of points from [a, b],
when we are interested in the integration process of this function. For
instance, if
½
1, x ∈ [0, 1)
f (x) = .
2 x ∈ [1, 2]
R2 R1 R2
Since 0 f (x)dx = 0 dx + 1 2dx = 1 + 2 = 3, (see example 26) we do
not care of the value of the function in its discontinuity point x = 1.
We can give a more general frame for all of the questions discussed
above, by introducing a new basic notion.
3. LEBESGUE CRITERION AND ITS APPLICATIONS 59
point for f. Finally we get that the set A of the discontinuity points of
|f | is contained in the set B of the discontinuity points of f. Since f is
Riemann integrable, the L-measure of B is zero. Thus the L-measure
of A is also zero (see example 28).
(n)
Let now {∆n } be a sequence of divisions of [a, b], ∆n : a < x1 <
(n) (n) (n)
x2 < ... < xkn = b, such that k∆n k → 0. Then S(f ; ∆n ; {ξ i }) →
(n)
I(f ) and S(|f | ; ∆n ; {ξ i }) → I(|f |). But
¯ n ¯
¯ ¯ ¯X ¯
¯ (n) ¯ ¯ (n) (n) (n) ¯
¯S(f ; ∆n ; {ξ i })¯ = ¯ f (ξ i )(xi − xi−1 )¯ ≤
¯ ¯
i=1
n ¯
X ¯
¯ (n) ¯ (n) (n) (n)
≤ ¯ i ¯ (xi − xi−1 ) = S(|f | ; ∆n ; {ξ i }).
f (ξ )
i=1
¯R ¯ R
¯ b ¯ b
So |I(f )| ≤ I(|f |), i.e. ¯ a f (x)dx¯ ≤ a |f (x)| dx. ¤
In the following we put together some basic properties of the Rie-
mann integrable functions space Int[a, b], defined on a fixed interval
[a, b].
Theorem 21. Let Int[a, b] be the set of all Riemann integrable
functions on [a, b].
a) Then Int[a, b] is a vector subspace of the vector space of all func-
tions defined on [a, b]. Moreover, if f, g ∈ Int[a, b] and α, β ∈ R, then
Z b Z b Z b
(3.1) [αf (x) + βg(x)]dx = α f (x)dx + β g(x)dx.
a a a
b) If f, g ∈ Int[a, b] and f (x) ≤ g(x) for any x ∈ [a, b], then
Rb Rb Rb
f (x)dx ≤ a g(x)dx. In particular, if f (x) ≥ 0, then a f (x)dx ≥ 0.
a Rb
In general, a f (x)dx may be zero but f (x) 6= 0 for at least one point
x.
c) If c ∈ (a, b) and if f ∈ Int[a, c] and if f ∈ Int[c, b], then f ∈
Int[a, b] and
Z b Z c Z b
(3.2) f (x)dx = f (x)dx + f (x)dx
a a c
R Rb Rd
If [a, b]∩[c, d] = ∅, we define [a,b]∪[c,d] f (x)dx = a f (x)dx+ c f (x)dx.
Ra Ra Rb
We also put a f (x)dx = 0 and b f (x)dx = − a f (x)dx. With this
notation, let l, m, n be three arbitrary real numbers in the interval [a, b],
then for f : [a, b] → R, integrable, one has:
Z n Z m Z n
(3.3) f (x)dx = f (x)dx + f (x)dx
l l m
62 2. DEFINITE INTEGRALS
Rb
d) Let g : [a, b] → [0, ∞) be a continuous function such that a
g(x)dx =
0. Then g(x) = 0 for any x in [a, b].
Proof. a) Let z be a continuity point for f and for g, where f, g ∈
Int[a, b] and let α, β be two arbitrary real numbers. Then we know
that z is also a continuity point for αf + βg, where (αf + βg)(x) =
αf (x)+βg(x) for any x ∈ [a, b]. Thus, the discontinuity set of points for
αf +βg is a subset of the union of the set of discontinuity points of f and
the set of discontinuity points of g. Since these last two sets have the
L-measure zero, we see that the L-measure of the set of discontinuity
points of αf + βg is also zero. Now, since f and g are bounded, it is
easy to see that αf + βg is also bounded. Let us use now theorem 17
and find that αf + βg is integrable. Formula (3.1) is true for Riemann
sums, so it is also true for their limits.
(n) (n) (n)
b) Let {∆n }, ∆n : a < x1 < x2 < ... < xkn = b, be a sequence of
(n) (n) (n)
divisions with k∆n k → 0 and let ξ i ∈ [xi−1 , xi ], i = 1, 2, ..., kn , be a
corresponding sequence of marking points of it. Since f (x) ≤ g(x) one
has that
(n)
X
n
(n) (n) (n)
S(f ; ∆n ; {ξ i }) = f (ξ i )(xi − xi−1 ) ≤
i=1
X
n
(n) (n) (n) (n)
≤ g(ξ i )(xi − xi−1 ) = S(g; ∆n ; {ξ i }).
i=1
Rb Rb
Taking limits here when n → ∞, we get that a f (x)dx ≤ a g(x)dx.
c) Since the discontinuity set of f on [a, b] is the union of the dis-
continuity sets of f on [a, c] and of f on [c, b] respectively, we see
that the L-measure of the first one is zero, i.e. f is integrable on
[a, b], if it is integrable on [a, c] and on [c, b] respectively. Let {∆0n },
0(n) 0(n) 0(n)
∆0n : a < x1 < x2 < ... < xkn0 = c, be a sequence of divisions
0(n) 0(n) 0(n)
of [a, c], with k∆0n k → 0 and let ξ i ∈ [xi−1 , xi ], i = 1, 2, ..., kn0 ,
be a corresponding sequence of marking points of it. Let also {∆00n },
00(n) 00(n) 00(n)
∆00n : a < x1 < x2 < ... < xkn00 = c, be a sequence of divisions
00(n) 00(n) 00(n)
of [c, b], with k∆00n k → 0 and let ξ i ∈ [xi−1 , xi ], i = 1, 2, ..., kn ,
be a corresponding sequence of marking points of it. We see that
(n) 0(n) 00(n)
∆n = ∆0n ∪ ∆00n is a division of [a, b] with {ξ i } = {ξ i } ∪ {ξ i } as a
corresponding set of marking points. Since k∆n k → 0, and because
(n) 0(n) 00(n)
S(f ; ∆n ; {ξ i }) = S(f ; ∆0n ; {ξ i }) + S(f ; ∆00n ; {ξ i }),
Rb Rc Rb
one has that a f (x)dx = a f (x)dx + c f (x)dx. Let us use now this
Ra Rb
last equality and the definition b f (x)dx = − a f (x)dx for a < b, in
4. MEAN THEOREM. NEWTON-LEIBNIZ FORMULA 63
order to prove (3.3). Assume for instance that m < n < l. Then, from
(3.2) we get:
Z l Z n Z l
f (x)dx = f (x)dx + f (x)dx.
m m n
So Z Z Z Z
n l n l
f (x)dx = − f (x)dx = f (x)dx − f (x)dx,
l n m m
Rn Rm Rn
or l f (x)dx = l f (x)dx + m f (x)dx.
d) Assume that there is a c ∈ [a, b] such that g(c) > 0 (if g(x) < 0,
the reasoning is completely analogous). Since g is continuous, there is
an entire interval [c − ε, c + ε] ⊂ [a, b], ε > 0, such that g(x) > 0 for
any x ∈ [c − ε, c + ε] (see [Po], theorem 34). Moreover, there is a point
x0 ∈ [c − ε, c + ε] with g(x0 ) = inf g(x) (see [Po], theorem 32).
x∈[c−ε,c+ε]
Thus, g(x0 ) > 0. But
Z b Z c−ε Z c+ε Z b
g(x)dx = g(x)dx+ g(x)dx+ g(x)dx ≥ 2ε·g(x0 ) > 0,
a a c−ε c+ε
which contradicts the hypothesis. Hence, g(x) = 0 for any x ∈ [a, b].
¤
Remark 13. We say that a number r ∈ R can be well approximated
with elements from a given subset A (the "approximation set") of R if
for any small real number ε > 0 one can find an element aε of A
such that |r − aε | < ε. For instance, any real number r can be well
approximated with rational numbers (here A is Q, the set or rational
numbers). Let now f : [a, b] → R be a Riemann integrable function.
Rb
Then I(f ) = a f (x)dx can be well approximated by an element of the
set of all Riemann sums of the form S(f ; ∆; {ξ i }) (see definition 3).
In particular, if g(x) = 1 for any x ∈ [a, b], we get the classical "mean
formula":
Z b
(4.2) f (x) dx = f (ξ)(b − a).
a
This last formula says that the trapezoidal area [AabB] (see Fig.6)
is equal to the area of the rectangle with the base [a, b] and the height
equal to the ordinate of f at a point ξ ∈ [a, b].
Figure 6
Since g(x) is continuous and not identical to zero, one has that
Z b
g(x) dx > 0
a
Rb
(see theorem 21), so we can divide the last inequalities by a
g(x) dx :
Rb
f (x)g(x) dx
m ≤ aR b ≤ M.
a
g(x) dx
4. MEAN THEOREM. NEWTON-LEIBNIZ FORMULA 65
We use now Darboux theorem (see [Po], Th. 33) Ufor the function f
b
f (x)g(x) dx
(f ([a, b]) = [m, M]) and find a ξ ∈ [a, b] such that a
Ub = f (ξ),
a g(x) dx
i.e. formula (4.1). ¤
Example 29. R 100Let us use the classical mean formula in order to
−x2
prove that I = 10 e dx < 2193 , so I is a very small number. Indeed,
2
from formula (4.2) we find that I = 90e−ξ , where ξ ∈ [10, 100]. But
2 27
the biggest value of e−x is e−100 < 2−100 , so I < 290 1
100 < 2100 = 293 .
But, using the Lagrange formula (see [Po], Ch.4, Cor.5) on the interval
[xi−1 , xi ], we find ci ∈ [xi−1 , xi ] such that F (xi ) − F (xi−1 ) = f (ci )(xi −
xi−1 ). Thus,
Xn
f (ci )(xi − xi−1 ) = F (b) − F (a).
i=1
¤
R1
Example 30. Let us compute 0 x2 dx. To use Newton-Leibniz for-
R 3
mula we need a primitive function for f (x) = x2 . Since x2 dx = x3 ,
R1 ¯1
3¯
one has that 0 x2 dx = x3 ¯ = 13 − 03 = 13 .
0
Ux 2
e−t dt
Example 31. Let us compute lim 0
x
. Since we have the non-
x→0
0
determinate case we can apply l’Hôspital rule. But the function
,
R x −t2 0
2
G(x) = 0 e dt is a primitive of g(x) = e−x (see theorem 23). So
2
G0 (x) = g(x) and our limit is equal to lim e−x = 1.
x→0
In this last integral we can make the change of variable tan z = t and
obtain:
Z π Z ∞
2 1 1
4 2 dz = 4 dt =
0 1 + sin z 0 1 + 2t2
Z ∞ ³√ ´ ¯
4 1 4 √ ¯∞ √
√ ¡√ ¢ d 2t = √ arctan 2t¯¯ = 2π,
2 0 1 + 2t 2 2 0
t2
because sin2 z = 1+t2
, 1
z = arctan t and dz = 1+t π
2 dt for z ∈ [0, 2 ). Here
Z ∞ Z A
1 def 1
2
dt = lim dt,
0 1 + 2t A→∞ 0 1 + 2t2
(q)
one has that m(Eext ) ≤ 23q1−q + 2q1+1 → 0, when q → ∞. Hence the
measure of C exists and it is equal to zero. An easy example of a
nonmeasurable subset of R is the set F = [0, 1] ∩ Q. Since F contains
no interior points w.r.t. R, it cannot contain a nontrivial interval. So,
if F were measurable, its measure would be zero. But this is not true
(q) (q)
because we cannot construct a sequence (Eext )q , where each Eext is a
finite union of intervals, two of which having in common at most one
(q) (q)
point, F ⊂ Eext for each q and m(Eext ) → 0, where q → 0. Indeed,
(q) (q)
in these conditions [0, 1] ⊂ Eext for all q = 1, 2, ..., thus m(Eext ) ≥ 1
(q)
and so the sequence (Eext )q cannot tend to 0. In the same manner one
can prove that if a subset A is measurable in R and has measure zero,
then it cannot have interior points. Conversely, if it is measurable and
has no interior points, then its measure must be zero (Why?). The
measure of a measurable subset A of R is simply called its length,
l(A). What is the connection between the notion of a subset of R of
Lebesgues measure zero (see Definition 5) and the notion of a subset
of R of measure zero? It is clear that a subset of R of measure zero
(it is enough to be measurable!) must be bounded. It is also clear
that a subset of R of measure zero has also the Lebesgue measure zero
(look carefully to both definitions!). The set N∗ = {1, 2, ..., n, ...} can
be covered by the disjoint union
1 1
∪∞
n=1 [n − ,n + ]
2ns+1 2ns+1
for any s ≥ 1. Since the sum of the lengths of all intervals of this union
is
X ∞
1 1 1 1
ns
= s· 1 = s
n=1
2 2 1 − 2s 2 −1
and since 2s1−1 → 0, when s → ∞, we see that the Lebesgue measure
of N∗ is zero and it is not bounded, i.e. it cannot be measurable, in
particular it cannot have the measure zero.
Another interesting example of a zero Jordan measure set of R is
the famous Cantor set. Take the interval A0 = [0, 1], divide it into 3
equal subintervals, get out the middle open subinterval (1/3, 2/3) and
obtain
A1 = [0, 1/3] ∪ [2/3, 1] ⊂ A0 .
Take now each of the both disjoint intervals of A1 and proceed in the
same way. We obtain
A2 = [0, 1/32 ] ∪ [2/32 ] ∪ [2/3, 7/32 ] ∪ [8/32 , 1] ⊂ A1 ⊂ A0 .
5. THE MEASURE OF A FIGURE IN Rn 71
Figure 7
Indeed, the area of the parallelogram [EF GH] is equal to the area
of the rectangle [HH 0 JG] = l(EF ) · l(HH 0 ). The area of the triangle
[KLM] is a half from the area of the parallelogram [KLNM]. To find
the area of a disc with centre at W and of radius R, we approximate
the disc with the surface bounded by a regular polygon with n sides
inscribed in the boundary circle of this disc. Now, the angle α, in
radians (see Fig.7), is equal to 2π 2n
= πn . So the area of the polygon is
equal to n · 2R sin πn · R cos πn · 12 . Hence, the area of the disc is
π sin πn
lim R2 cos · π · π = πR2 .
n→∞ n n
The length of the boundary circle of the above disc is
π sin π
lim n · 2R sin = 2πR lim π n = 2πR.
n→∞ n n→∞
n
72 2. DEFINITE INTEGRALS
Figure 8
The intersection of two rectangles which belong to an elementary
figure has no interior points w.r.t. R2 if and only if either it is empty,
it is a point or a segment of a line parallel to one of the axes. For a
plane elementary figure see Fig.9.
In Fig.10 we see how to cover "from interior" with an elementary
figure a general plane figure A.
In Fig.11 we see how to cover "from exterior" with an elementary
figure another general figure A.
In Fig.12 we see the simultaneous process of approximation "from
interior" and "from exterior" with elementary figures Ei and Eo respec-
tively, a general plane figure B.
The measure of a measurable figure A in R2 (a plane figure!) is
simply called its area σ(A). The areas of a finite union of points or of
a finite union of segments of lines are zero (Why?). The same is true
for a finite union of curves of class C 1 (piecewise smooth curves! ).
Remark 14. To define the area σ(A) of a measurable subset A of
R2 we used the above finite unions of rectangles as elementary figures.
Since any rectangle is obvious a union of two triangles (as surfaces!)
with their intersections a segment of zero areas and since any triangle
5. THE MEASURE OF A FIGURE IN Rn 73
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Since
⎧
⎨ −x − 1, x ∈ [−2, −1),
|f (x)| = −x3 , x ∈ [−1, 0),
⎩ x3 , x ∈ [0, 1],
one has
Z 1 Z −1
area(Df ) = |f (x)| dx = (−x − 1)dx +
−2 −2
Z 0 Z 1
3
(−x )dx + x3 dx
−1 0
µ
¶¯−1 ¯
4 ¯0
¯
4 ¯1
x2 ¯ x x
= − + x ¯¯ − ¯¯ + ¯¯
2 −2 4 −1 4 0
= 1.
Figure 15
But whenever on a subinterval [c, d] ⊂ [a, b], g(x) ≥ f (x), x ∈ [c, d],
we compute the area "between g and f " by the formula
Z d Z d
[g(x) − f (x)]dx = |f (x) − g(x)| dx.
c c
Indeed,
Z c Z d
areaf,g = [g(x) − f (x)]dx + [f (x) − g(x)]dx+
a c
Z e Z b Z b
+ [g(x) − f (x)]dx + [f (x) − g(x)]dx = |f (x) − g(x)| dx.
d e a
Example 36. Let us compute the area between f (x) = sin x and
g(x) = sin 2x, x ∈ [0, π]. The best way is to look at the graphics of f
and g in Fig.17.
Since the intersection points of these graphics are x = 0, x = π3 and
x = π, one can write:
Z π
areaf,g = |sin x − sin 2x| dx.
0
80 2. DEFINITE INTEGRALS
Figure 16
Figure 17
Since a continuous function keeps the same sign between two consec-
utive zeros of it (Why?-see Darboux theorem in [Po]), the sign of the
function h(x) = sin x − sin 2x is the sign of h( π4 ) = √12 − 1 < 0 on
the interval [0, π3 ]. The sign of h(x) on the interval [ π3 , π] is the sign of
h( 3π
4
) = √12 + 1 > 0. Thus,
Z π Z π
3
areaf,g = (sin 2x − sin x)dx + (sin x − sin 2x)dx =
π
0 3
¯π ¯π
cos 2x ¯¯ 3 π
π cos 2x ¯¯
− + cos x |0 − cos x | π +
3
=
2 ¯0 3 2 ¯π
3
6. AREAS OF PLANE FIGURES BOUNDED BY GRAPHICS OF FUNCTIONS81
1 1 1 1 1 1 11
+ + −1+1+ + + = .
4 2 2 2 2 2 4
Assume now that we have a plane parametric curve
½
x = x(t)
(Γ) : , x ∈ [a, b],
y = y(t)
such that x(t) is an increasing function of class C 1 and y(t) is a non-
negative continuous function (see Fig.18).
Figure 18
Then the area of the domain D(Γ) , bounded by (Γ), the segment
[x(a), x(b)] and the lines x = x(a), x = x(b), can be approximated by
Riemann’s sums of the form:
X
n
(6.5) S(Γ) (∆; (ξ i )) = y(ξ i ) (x(ti ) − x(ti−1 )) ,
i=1
where ∆ : a = t0 < t1 < ... < b = tn is a division of the interval [a, b]
and (ξ i ), i = 1, 2, ..., n, ξ i is an arbitrary set of fixed marking points
for ∆. One can use Lagrange’s formula for function x(t) restricted to
[ti−1 , ti ] and find:
x(ti ) − x(ti−1 ) = x0 (η i )(ti − ti−1 )
for an η i ∈ (ti−1 , ti ). When k∆k → 0, η i and ξ i become closer and
closer. Since x0 (t) is continuous, x0 (η i ) and x0 (ξ i ) become closer and
closer. So the area above can be well approximated by sums of the
form:
X
n
∗
S(Γ) (∆; (ξ i )) = y(ξ i )x0 (ξ i ) (ti − ti−1 ) .
i=1
82 2. DEFINITE INTEGRALS
But these last sums are Riemann’s sums for a new function f (t) =
y(t)x0 (t), t ∈ [a, b]. Hence the area above can be computed by the
formula
Z b
(6.6) area(D(Γ) ) = y(t)x0 (t)dt.
a
If x(t) is increasing or decreasing almost everywhere (it is increasing
or decreasing on each subinterval of a fixed division a = c0 < c1 < ... <
ck = b), then obviously formula (6.6) must be substituted with a more
general formula
Z b
(6.7) area(D(Γ) ) = y(t) |x0 (t)| dt.
a
The same formula works if in addition to this last generalization, x(t)
is a smooth piecewise function on [a, b] and (or) y(t) is piecewise con-
tinuous on the same interval.
Example 37. Let us compute the area bounded by the axis Ox and
the arc of the cycloid
½
x = a(t − sin t)
(Γ) : , t ∈ [0, 2π],
y = a(1 − cos t)
where a > 0 is a parameter. Since x0 (t) = a(1 − cos t) ≥ 0, we can
apply formula (6.6) and find
Z 2π Z 2π
2 2 2
area = a (1 − cos t) dt = a (1 − 2 cos t + cos2 t)dt =
0 0
Z 2π
a2
= 2πa2 + 0 + (1 + cos 2t)dt = 3πa2 .
2 0
Figure 19
of the disc of radius ρ(ξ i ) and centre O, bounded by the arc (θi−1 θi ) .
A very known elementary formula says that this last area is equal to
1
2
[ρ(ξ i )]2 (θi − θi−1 ), where θi − θi−1 is measured in radians. This last
formula is in fact the area of a triangle of basis ρ(ξ i )(θi − θi−1 ) (the real
length of the arc (θi−1 θi ) of that circle) and height ρ(ξ i ). So the total
area can be well approximated by the following sum
X
n
1
Sρ(θ) (∆; {ξ i }) = ρ2 (ξ i )(θi − θi−1 ).
i=1
2
But this last sum is exactly a Riemann sum for the function θ → 12 ρ2 (θ),
θ ∈ [α, β]. Hence our area is equal to the corresponding Riemann inte-
gral
Z
1 β 2
(6.8) area = ρ (θ)dθ.
2 α
Example 38. The cardioid is the curve ρ = ρ(θ) = a(1 + cos θ),
θ ∈ [0, 2π) and a > 0 is a parameter (see Fig.20). Let us use formula
(6.8) in order to compute the area bounded by this cardioid:
Z ∙ Z 2π ¸
1 2π 2 2 a2 2
area = a (1 + cos θ) dθ = 2π + cos θdθ =
2 0 2 0
Z
2 a2 2π 3πa2
= πa + (1 + cos 2θ)dθ = .
4 0 2
Figure 20
−−→
Since the ray OM, M(ρ(θ), θ) covers the double hatched region I two
times the area of the hatched region is:
Z 5π Z π
1 2
2 1 2 2 93π 3
area = 9θ dθ − 9θ dθ = .
2 0 2 0 4
Figure 21
Example 40. The graphic of the curve ρ = ρ(θ) = a sin 3θ, where
θ ∈ [0, 2π] has the form of a clover with three leaflets (see Fig.22).
Since sin 3θ must be greater or equal to zero, we have that
π 2π 4π 5π
θ ∈ [0, ] ∪ [ , π] ∪ [ , ].
3 3 3 3
7. THE VOLUME OF A ROTATIONAL SOLID 85
Figure 22
Figure 23
a Riemann sum for the function πf 2 , the division ∆ and the set of
marking points (ξ i ). So
Z b
(7.3) vol(D) = π f 2 (x)dx.
a
Formula (7.2) says that we can well approximate the volume of the ro-
tational solid D with the sum of volumes of the cylinders Ci , generated
by the rotation of the line y(x) = f (ξ i ) for any x ∈ [xi−1 , xi ], around
Ox-axis (see Fig.24).
Example 41. Let us find the volume of a ball of radius R > 0, in
R3 .
Such√a ball is the rotational solid generated by the rotation of the
arc y = R2 − x2 , x ∈ [−R, R] of the circle of radius R, x2 + y 2 = R2 ,
around Ox-axis. Thus,
Z R Z R
2 2
vol = π (R − x )dx = 2π (R2 − x2 )dx =
−R 0
8. THE LENGTH OF A CURVE IN R3 87
Figure 24
µ ¶¯R
2 x3 ¯¯ 4πR3
= 2π R x − = .
3 ¯0 3
for the length of (Γ). This number l(Γ) defined above is unique. Let
us estimate now l(Γ).
Let us consider a parametric representation of a smooth curve (Γ)
(see (8.1)), a division ∆ : a = t0 < t1 < ... < tn = b of the interval [a, b]
and a set of marking points (ξ i )i , ξ i ∈ [ti−1 , ti ] for any i = 1, 2, ..., n.
By definition 8 we can approximate the length l(Γ) with
X n ° °
°−−−−−→°
(8.2) S= °Mi−1 Mi ° =
i=1
X
n
p
= [x(ti ) − x(ti−1 )]2 + [y(ti ) − y(ti−1 )]2 + [z(ti ) − z(ti−1 )]2 .
i=1
(see Fig. 25).
Figure 25
Since x(t), y(t) and z(t) are of class C 1 , applying Lagrange formula
to them on each interval [ti−1 , ti ], we get
x(ti ) − x(ti−1 ) = x0 (ai )(ti − ti−1 ) ≈ x0 (ξ i )(ti − ti−1 ),
y(ti ) − y(ti−1 ) = y 0 (bi )(ti − ti−1 ) ≈ y 0 (ξ i )(ti − ti−1 ),
z(ti ) − z(ti−1 ) = z 0 (ci )(ti − ti−1 ) ≈ z 0 (ξ i )(ti − ti−1 ),
where ai , bi , ci ∈ [ti−1 , ti ] for each i = 1, 2, ..., n. Thus, S can be well
approximated (when k∆k → 0) with
X n
p
S1 = x0 (ξ i )2 + y 0 (ξ i )2 + z 0 (ξ i )2 (ti − ti−1 ).
i=1
8. THE LENGTH OF A CURVE IN R3 89
ε
≤
· (b − a) = ε.
b−a
But this sum S1 is a Riemann sum for the function
p
f (t) = x0 (t)2 + y 0 (t)2 + z 0 (t)2 .
So our curve (Γ) is rectifiable and its length can be computed by using
the formula
Z bp
(8.3) l(Γ) = x0 (t)2 + y 0 (t)2 + z 0 (t)2 dt.
a
p
The expression ds = x0 (t)2 + y 0 (t)2 + z 0 (t)2 dt is called the element
of length on the arc Γ, i.e. it is "the limit" of the length of the arc
M\i−1 Mi on Γ, when the distance between Mi−1 Mi is small enough.
Let us compute the length of the astroide
½
x = a cos3 t
, t ∈ [0, 2π], a > 0
y = a sin3 t
This curve is a plane curve, so z(t) = 0 in the above formula. We see
that the Cartesian form of the parametric equation of the astroide is
2 2 2
x 3 + y 3 = a 3 . Thus the curve is symmetric relative to Ox and Oy axes.
Hence its length is
Z π
p Z π
p
2 2
2 4
l=4 2 4 2 2
9a cos t sin t + 9a sin t cos tdt = 12a cos2 t sin2 tdt =
0 0
Z π Z π
π
2 2
= 12a |cos t sin t| dt = 6a sin 2tdt = −3a cos 2t|02 = 6a.
0 0
90 2. DEFINITE INTEGRALS
Figure 26
b−a
namely equidistant divisions. Let n be large enough and let h = n
,
"the increment" of the division
∆n = a = x0 < a + h = x1 < ... < a + ih = xi < ... < xn = a + nh = b.
This division is an example of an equidistant division because xi −
xi−1 = h, a constant number. Let us take ξ i = xi−12+xi , the midpoint of
the interval [xi−1 , xi ]. Thus,
a + (i − 1)h + a + ih 2i − 1
ξi = =a+ h.
2 2
So we can use in practice the following approximation:
Z b µ ¶
b−aX
n
2i − 1
(9.1) f (x)dx ≈ R(f ; n) = f a+ h .
a n i=1 2
RThis
b
formula is called the rectangles formula. In Fig.27 the integral
a
f (x)dx is just the area of the plane surface bounded by the graphic
of f (x), when x ∈ [a, b], the Ox-axis and the vertical lines x = a, y = b.
The approximate R(f ; n) from formula (9.1) is exactly the hatched area
of Fig.27, i.e. the sum of the areas of all the rectangles Di , where Di
has as a basis the segment [xi−1 , xi ] and as height f (ξ i ), ξ i being the
midpoint of [xi−1 , xi ].
Figure 27
Figure 28
Let us denote by M2 = sup |f 00 (x)| and let us apply Taylor formula
x∈[a,b]
of order 2 around the point ci , i.e.
f 0 (ci ) f 00 (c)
(9.4) f (x) = f (ci ) + (x − ci ) + (x − ci )2 ,
1! 2!
where ci is a point in the interval [xi−1 , xi ]. Thus,
¯Z xi µ ¶ ¯
¯ x + x ¯
erri = ¯¯ (xi − xi−1 )¯¯ =
i−1 i
f (x)dx − f
xi−1 2
9. APPROXIMATE COMPUTATION OF DEFINITE INTEGRALS. 93
¯Z xi ¯ ¯Z ¯
¯ ¯ |f 0 (ci )| ¯ xi ¯
=¯ ¯ ¯
[f (x) − f (ci )]dx¯ ≤ ¯ (x − c )dx ¯+
1! ¯ i ¯
¯Z xxii−1 ¯ ∙
xi−1
¸
M2 ¯¯ ¯ M (x − c )3
(x − c )3
(x − ci ) dx¯¯ = 0 +
2 2 i i i−1 i
+ − =
2 ¯ xi−1 2 3 3
" x +x x +x
#
M2 (xi − i−12 i )3 (xi−1 − i−12 i )3 h3 M2
= − = M2 = 3
(b − a)3 .
2 3 3 24 24n
The global error
¯ ¯ ¯ n Z ¯
¯ X
n ¯ ¯X xi ¯
¯ ¯ ¯ ¯
err = ¯I − f (ci )(xi − xi−1 )¯ = ¯ (f (x) − f (ci )) dx¯ ≤
¯ ¯ ¯ xi−1 ¯
i=1 i=1
X n
M2 M2
≤ erri ≤ n · 3
(b − a)3 = (b − a)3 .
i=1
24n 24n2
Thus
M2
(9.5) err ≤ (b − a)3 ,
24n2
where M2 = sup |f 00 (x)| .
x∈[a,b]
Let us come back to example 43 and evaluate ³ the´ error we have
R 1 −x2 −x2
made by approximating I = 0 e dx with R e ; 5 . Let us evalu-
ate M2 :
2 2
f 0 (x) = −2xe−x ; f 00 (x) = (−2 + 4x2 )e−x .
Thus, |f 00 (x)| ≤ 2 when x ∈ [0, 1] and
2 1 1
err ≤ · 13 = < .
24 · 25 300 100
Hence, the sum from (9.2) approximates our integral with at least 2
exact decimals.
The general idea above is to substitute the original function f :
[a, b] → R with another one, called an interpolating function, usually
more elementary then the initial one. In our case of rectangles ap-
proximation, we substituted f with fe such that fe(x) =R f (ci ) for any
1 2
x ∈ [xi−1 , xi ), i = 1, 2, ..., n. Then we approximated I = 0 e−x dx with
Z b Xn Z xi Xn
e
f (x)dx = f (ci )dx = f (ci )(xi − xi−1 ) = R(f, n).
a i=1 xi−1 i=1
Figure 29
b−a
2n
[f (xi−1 ) + f (xi )] . Thus,
X
n
b−a
(9.6) I≈ [f (xi−1 ) + f (xi )] =
i=1
2n
" #
b−a Xn
def
= f (a) + f (b) + 2 f (xi ) = T (f, n).
2n i=1
This last formula is called the trapezoids formula. One can also estimate
the error err = |I − T (f, n)| in this case:
M2
(9.7) (b − a)3 ,
err ≤
12n2
where M2 = sup |f 00 (x)| . For a proof of this last inequality see for
x∈[a,b]
instance [GG], pag.139.
9. APPROXIMATE COMPUTATION OF DEFINITE INTEGRALS. 95
Even our feeling says that trapezoids formula gives a better approx-
imation than the rectangles formula, the mathematical estimation of
errors (9.5) and (9.7) say contrary, namely, the estimation of error in
the case of the rectangles formula is two times smaller than the estima-
tion in the case of the trapezoids formula. In practice, sometimes the
error in the case of the trapezoids approximation formula can be less
than that one in which we use rectangle formula (for the same equidis-
tant division). Sometimes it is greater! (by drawing, find out some
examples!). Since both R(f, n) and T (f, n) are convergent to I, when
n → ∞, for n large enough one can use either rectangles or trapezoids
formula.
In order to obtain "better" approximation formulas we need to "in-
terpolate" our integrand function f (x) by a polynomial P (x) of a fixed
degree n. This means to fix n + 1 distinct points x0 < x1 < ... < xn
in [a, b] and to find a polynomial P (x) of degree n such that P (xi ) =
f (xi ) for any i = 0, 1, ..., n. We shall see that this last polynomial
is unique and it is called the interpolation polynomial of f at the
nodes {x0 , x1 , ..., xn }. In practice, we usually do not know the ana-
lytical expression of the function f (x). In fact we measure some values
{y0 , y1 , ..., yn } of it at a given finite number of points x0 < x1 < ... < xn .
Thus, f (xi ) = yi for i = 0, 1, ..., n. Having such an interpolation poly-
nomial we can force the approximation f (x) ≈ P (x), x ∈ [a, b]. One can
prove that if the number of nodes is greater and greater and if they are
"uniformly" distributed (for instance if we use equidistant divisions of
[a, b]), then the error kf − P k = sup {|f (x) − P (x)|} becomes smaller
x∈[a,b]
and smaller (see [GG] for instance). Let us use this last approximation
in order to find an approximation of the integral
Z b Z b
f (x)dx ≈ P (x)dx.
a a
This last integral can be easily computed even the degree of P is very
large (we can derive an elementary formula which is a function of the
coefficients of P and of a and b). Thus, the approximate computation
Rb
of a f (x)dx reduces to an easy task.
Let us reformulate our general problem. Given n + 1 nodes (points,
numbers, etc.) x0 < x1 < ... < xn and a fixed arbitrary set of n + 1
real numbers {y0 , y1 , ..., yn }, let us find a polynomial Ln (x) (here L is
from Lagrange) of degree n such that
Ln (xi ) = yi , i = 0, 1, ..., n,
96 2. DEFINITE INTEGRALS
Figure 30
Such a polynomial Ln (x) is called a Lagrange polynomial corre-
sponding to the nodes x0 < x1 < ... < xn and to {y0 , y1 , ..., yn }.
Theorem 27. Given x0 < x1 < ... < xn and {y0 , y1 , ..., yn }, there
is a unique Lagrange polynomial Ln (x) of degree n which is passing
through the points Mi (xi , yi ), i = 0, 1, ..., n.
Proof. Any polynomial Ln (x) of degree n can be represented as
Ln (x) = a0 + a1 x + ... + an xn ,
where a0 , a1 , ..., an are unknown coefficients. To determine Ln (x) means
to find these coefficients a0 , a1 , ..., an . Conditions Ln (xi ) = yi , i =
0, 1, ..., n can be also written as a linear system in variables a0 , a1 , ..., an :
⎧
⎪
⎪ a0 + a1 x0 + ... + an xn0 = y0
⎪
⎪
⎨ a0 + a1 x1 + ... + an xn1 = y1
............................ .
⎪
⎪ a + a x + ... + a x n
= y
⎪
⎪ 0 1 n−1 n n−1 n−1
⎩ a0 + a1 xn + ... + an xnn = yn
The determinant of the system is a Vandermonde determinant equal
to Y
(xj − xi ), which is not zero because xi 6= xj for any i 6= j.
0≤i<j≤n
Thus the solution exists and it is unique. ¤
9. APPROXIMATE COMPUTATION OF DEFINITE INTEGRALS. 97
{ω0 (x), ω 1 (x), ..., ω n (x)} is linear independent. Indeed, let us take a
linear combination of them:
(9.11) λ0 ω0 (x) + λ1 ω1 (x) + ... + λn ωn (x) = 0,
for any x ∈ R. Let us put x = x0 in (9.11). Since ω 0 (x0 ) 6= 0
(x0 , x1 , ..., xn are distinct), ω 1 (x0 ) = 0, ..., ω n (x0 ) = 0, we get that
λ0 = 0. If we put now x = x1 in 9.11, we find λ1 = 0, etc. Finally we find
that all λ0 s are zero, i.e. the set {ω 0 (x), ω 1 (x), ..., ω n (x)} is linear inde-
pendent in Pn . Having n + 1 elements and dimension being n + 1, any
other polynomial Q cannot be outside the linear subspace generated by
{ω0 (x), ω 1 (x), ..., ω n (x)}. If we could find such a polynomial, we could
enlarge the number of linear independent elements to n+2 which is not
possible because the dimension is exactly n + 1 (see a course in Linear
Algebra, for instance [Po1] or [Po2]). So {ω 0 (x), ω1 (x), ..., ω n (x)} is
also a generating system for Pn , i.e. this set of polynomials is a basis
of Pn .
b) Since {ω 0 (x), ω 1 (x), ..., ω n (x)} is a basis, one can write Ln (x) as
(9.12) Ln (x) = μ0 ω0 (x) + μ1 ω 1 (x) + ... + μn ω n (x).
To compute μ0 , μ1 , ..., μn , we successively make x = x0 , x1 , ..., xn in
(9.12). For instance, if x = x0 , we get
y0 = Ln (x0 ) = μ0 ω 0 (x0 ),
because ω1 (x), ..., ω n (x) are zero at x0 . Thus, μ0 = ω0y(x0 0 ) . In the same
way we get that μ1 = ω1y(x1 1 ) , ..., μn = ωny(xn n ) , i.e. we just proved formula
(9.10). ¤
Example 44. (interpolation with parabolas) Let us interpolate three
points M1 (x1 , y1 ), M2 (x2 , y2 ), M3 (x3 , y3 ) with a parabola, i.e. with a
Lagrange polynomial of degree 2. Let us use formula (9.10) and find
y1 y2 y3
(9.13) L2 (x) = ω1 (x) + ω2 (x) + ω 3 (x),
ω1 (x1 ) ω 2 (x2 ) ω 3 (x3 )
where ω 1 (x) = (x − x2 )(x − x3 ), ω 2 (x) = (x − x1 )(x − x3 ) and ω3 (x) =
(x − x1 )(x − x2 ).
Exercise 5. Write L2 (x) for x1 = 0.1, x2 = 0.5, x3 = 0.9 and
y1 = 1, y2 = 1.5, y3 = 0.5. Draw the graphic of the resulting parabola!
Assume now that in Example 44 x2 is the midpoint of the segment
[x1 , x3 ]. Thus, x2 = x1 +x
2
3
. Let us denote by h the differences
x3 − x1
(9.14) x2 − x1 = x3 − x2 = = h.
2
9. APPROXIMATE COMPUTATION OF DEFINITE INTEGRALS. 99
Rx
Let us evaluate I = I[x1 , x3 ; y1 , y2 , y3 ] = x13 L2 (x)dx. Let us use now
formula (9.13) to compute this last integral. Since
(9.15)
Z x3 Z x3 Z x3 Z x3
ω 1 (x) ω 2 (x) ω 3 (x)
L2 (x)dx = y1 dx + y2 dx + y3 dx.
x1 x1 ω 1 (x1 ) x1 ω 2 (x2 ) x1 ω 3 (x1 )
and if one uses formula (9.16) to compute each term of this last sum,
one gets
Z b
b−a
f (x)dx ≈ [f (x0 ) + f (x2n )+
a 6n
(9.17)
+4(f (x1 ) + f (x3 ) + ... + f (x2n−1 )) + 2(f (x2 ) + f (x4 ) + ... + f (x2n−2 ))].
100 2. DEFINITE INTEGRALS
Figure 31
15. Find the area bounded by the curve: ρ = a(1+cos 2θ), θ ∈ [0, π]
(Draw it!).
16. Find the area bounded by the Bernoulli’s lemniscate: ρ2 =
a cos 2θ, θ ∈ [− π4 , π4 ] ∪ [ 3π
2
, 5π ].
4 4 √ √
17. Find the length of the curves: a) y = ln x, x ∈ [ 3, 8]; b)
x = 14 y 2 − 12 ln y, y ∈ [1, e].
18. Find the lengths½ of the following curves:
x = a(t − sin t),
a) the cycloid: , t ∈ [0, 2π], a > 0,
y = a(1 − cos t)
b) the cardioid: ρ = a(1 + cos θ), θ ∈ [0, 2π] and
2 2 2
c) the astroide: x 3 + y 3 = a 3 , a > 0.
19. Find the volume of the revolutionary ellipsoid determined by
2 2
the rotation of the ellipse xa2 + yb2 = 1, a) around Ox-axis; b) around
Oy-axis. R? p
20. Use the formula S = 2π ? y 1 + y 02 dx to find the area of
the parabolic mirror generated by the rotation of the arc of parabola
1
x = 16 ay 2 , y ∈ [0, 4a], around Ox-axis.
21. The curve y = a cosh xa (catenary), a > 0, x ∈ [0, a] rotates
around the Ox-axis and gives rise to a surface called catenoid. Find
the area of this catenoid.
22. Find the revolutionary areas obtained by the rotation around
the Ox-axis of the a) cycloid, b) cardioid and c) astroide from the
problem 18.
23. The circle x2 + (y − b)2 = a2 , b > a, rotates around Ox-axis.
Find the side area and the volume of the resulting solid (torus).
24. The parabola y = x2 , x ∈ [0, 1] rotates itself around the first
bisectrix y = x. Find the side area and the volume of the resulting solid
(Hint: use Riemann’s sums to find appropriate formulas or make first
of all a rotation of axes√ of 45◦ ).
R 3 √1+x2 R 2e
25. Compute: a) 1 2 dx; b) 1 |ln x − 1| dx;
R1 £¡ ¢x ¤ x
c) −1 max 13 , 3x dx;
Rπ
26. Let us denote by In the following definite integral 02 sinn xdx,
n = 1, 2, ... . Prove that In = n−1 I . Then compute I5 and I6 .
n n−2
10. PROBLEMS AND EXERCISES 103
27. Find the coordinates of the mass center of the plane domain
bounded by the arc of parabola y 2 = x, y ≥ 0, the Ox-axis and the
line x = 2.
28. Find the matrix of the momentum of inertia for the arc of the
helix: ⎧
⎨ x = a cos t
H := y = a sin t , a, b > 0, t ∈ [0, 2π].
⎩ z = bt
⎛ ⎞
Ixx Ixy Ixz R 2π
(Hint: I = ⎝ Ixy Iyy Iyz ⎠ , where Ixx = 0 (y 2 + z 2 )ds, Ixy =
Ixz Iyz Izz
R 2π 2
z ds, etc., i.e. the square of the distance of a point M(x, y, z) of H
0 p
to the Ox-axis, to the Oxy-plane, etc. Here ds = x02 + y 02 + z 02 dt is
the element of length on the curve p H.
29. Use thepformula ds = x02 + y 02 dt to deduce an analogous
formula, ds = ρ2 + ρ02 dθ, for a curve given in polar coordinates:
ρ = ρ(θ). Use this last formula to compute the length of the cardioid:
ρ = a(1 + cos θ), θ ∈ [0, 2π], where a > 0.
2
30. The parabola y = x4 , x ∈ [0, 3], rotates around Oy-axis. Find
the volume of the resultant solid.
31. Find the coordinates of the mass center of the figures bounded
by the curves:
2 2 √
a) xa2 + yb2 = 1, x, y ≥ 0, x = 0, y = 0; b) y = x2 , y = x, x ≥ 0.
32. Use the formulas (see also the chapter with line integrals of the
first type):
Z 2π p
L = x02 + y 02 dt,
0
R 2π p
x x02 + y 02 dt
xG = 0 ,
L
R 2π p
y x02 + y 02 dt
yG = 0
L
in order to compute the coordinates of the mass center of the line:
½
x = a(t − sin t)
, t ∈ [0, 2π).
y = a(1 − cos t)
33. Let M1 (0, 12 ), M2 (1, 1), M3 (2, 12 ), M4 (3, 1) and M5 (4, 12 ) be 5
points in the xOy-plane. Write the Lagrange’s polynomial for these
points and compute the area bounded by it, the lines x = 0, x = 4,
y = 0 and the graphic of this polynomial.
104 2. DEFINITE INTEGRALS
Remark 16. Since on the real line R we have two directions around
a given point b, on the left and on the right, we might be inclined to
generalize theorem 29 to the case of a function of many variables, say
of two variables, in the following sense. Let A be a nonempty domain
(open and connected, see [Po] for instance) of the xOy-plane and let
b = (b1 , b2 ) be a limit point of A. Let f : A → R be a function of
two variables defined on A with scalar real values. The temptation is
to say that if the limit along any straight line which is passing through
b exists and it does not depend on the direction of these lines (this
means that there exist a real number L such that if {Mn = (xn , yn )} is
a sequence of points on a line y − b2 = m(x − b1 ) for all slopes m, i.e.
yn − b2 = m(xn − b1 ) for any n, then f (xn , yn ) has a limit at b). But
this statement is wrong as follows from the example bellow. Let
½ y √
3 x+y , if y 6= − x
√ 3
f (x, y) = √ ,
0, if y = − x 3
−1 xn + yn −1
1 + xynn
(2.1) = lim tan = tan lim
xn →∞,yn →∞ 1 − xn yn xn →∞,yn →∞ 1 − yn
xn
Since xynn has no invariant limit whenever xn , yn are different sequences
going to ∞, the limit in formula (2.1) does not exists. Thus, we can-
not study our integral with this definition. In spite of all of this, the
principal value of the integral is equal to π! This is why tacitly such
integrals are computed as the principal values of them (if these ones
exist!).
Definition 10. Let f : [a, ∞) → R be a function defined on the
unbounded interval (infinite interval) [a, ∞) with real values such that
it is integrable on anyR subinterval of the form [a, x] ⊂ [a, ∞). If the
x
new function F (x) = a f (t)dt,R F : [a, ∞) → R has a limit L ∈ R at
∞
b = ∞, we say that the symbol a f (x)dx is convergent (or exists) and
its value is that L, i.e.
Z ∞ Z x
f (x)dx = lim f (t)dt,
a x→∞ a
if this last
R ∞ limit exists. If the limit does not exist, we say that the
integral a f (x)dx is divergent or that it does not exist. Such inte-
R∞ Rb R∞
grals like a f (x)dx, −∞ f (x)dx, or −∞ f (x)dx are called improper
(generalized) integrals of the first type.
R∞ 1
For instance, 0 1+x 2 dx is convergent because
Z x
1 π
lim 2
dt = lim [arctan x − arctan 0] = .
x→∞ 0 1 + t x→∞ 2
R∞ 1 π
Thus, 0 1+x2 dx = 2 .
Example 46. (Gabriel’s trumpet) Let us take a > 0 and f (x) = x1
defined on the infinite interval [a, ∞). Let us take A > a and let us
rotate the arc of the graphic of f restricted to [a, A] around Ox-axis.
The side surface of the solid obtained in this way is like a trumpet. Let
us make A go to ∞. The "infinite" trumpet obtained in this way is
called a Gabriel’s trumpet. The infinite area bounded by the graphic of
f (x), x ∈ [a, ∞), the lines x = a and the Ox-axis (see Fig.1)
is Z ∞ Z x
1 1
dx = lim dt = lim [ln x − ln a] = ∞,
a x x→∞ a t x→∞
so this last improper integral is divergent. Let us compute the volume
of the revolutionary solid bounded by the Gabriel’s trumpet:
Z ∞ Z x ∙ ¸
1 1 1 1 π
π 2
dx = π lim 2
dt = π lim − + = .
a x x→∞ a t x→∞ x a a
114 3. IMPROPER (GENERALIZED) INTEGRALS
Figure 1
Thus the volume of this infinite solid is finite, namely πa . Hence the last
improper integral of the first type is convergent. If a > 0 goes to zero,
the volume goes to ∞. Very strange! The area of a longitudinal central
section of the Gabriel’s trumpet is infinite and its volume is finite!
Theorem 32. (zero criterion) Let us consider R∞ the conditions and
notation from definition 10. If the integral a f (x)dx is convergent,
then there exists at least one sequence
R∞ {qn } , qn → ∞, such that f (qn ) →
0. In particular, if the integral a f (x)dx is convergent, and if lim f (x)
R∞ x→∞
exists, then it must be zero. Thus, if the integral a f (x)dx is conver-
gent, if f is continuous and if lim f (x) exists, then f must be bounded.
x→∞
Hence, if lim f (x) exists and it is not equal to zero, then the improper
R x→∞
∞
integral a f (x)dx is divergent.
Proof. Let n0 be the least natural number greater then a. Then
Z ∞ Z n0 ∞ Z n+1
X
f (x)dx = f (x)dx + f (x)dx.
a a n=n0 n
But Z Z
n+1 n+1
f (x)dx = f (qn )
dx = f (qn ),
n n
P R n+1
where qn ∈ [n, n+1] (mean theorem). Since the series ∞ n=n0 n f (x)dx
is convergent, its general term f (qn ) converges to zero, when n → ∞.
The other statements are simple consequences of the definition of a
limit of a function at ∞ ( lim f (x) = lim f (qn ) = 0 for any sequence
x→∞ n→∞
qn → ∞) and of the basic properties of sequences on a compact inter-
val (see [Po], Weierstrass theorem). Indeed, if f were not bounded,
there exists a sequence {xn }, xn ∈ [a, ∞) for n = 1, 2, ..., such that
2. IMPROPER INTEGRALS OF THE FIRST TYPE 115
f (xn ) → ∞ (say; you can easily put −∞ instead ∞). If {xn } itself is
bounded, there is a subsequence {xkn } of {xn } with xkn → x0 ∈ [a, ∞)
(Cesàro’s Lemma, a particular case of Weierstrass theorem, see [Po]).
The continuity of f implies that f (xkn ) → f (x0 ). Since {f (xkn )} is a
subsequence of {f (xn )}, then f (xkn ) → ∞. The uniqueness of the limit
of a sequence implies that f (x0 ) = ∞, a pure contradiction! Thus the
sequence {xn } is not bounded to ∞. Let {xtn } be a subsequence of
{xn } with xtn → ∞. Since lim f (x) = 0, one has that f (xtn ) → 0. But
x→∞
f (xtn ) → ∞, {xtn } being a subsequence of {xn } , thus we again obtain
a contradiction (0 = ∞!). Hence f cannot be unbounded. ¤
Here
R ∞ is an example of an unbounded function f (x) on [a, ∞) such
that a f (x)dx is convergent (contrary to a Riemann integral on a
finite and closed interval [a, b]!). Take N large enough such that N ≥ a
and define f : [a, ∞) → R, f (N + i) = N + i for any i = 0, 1, 2, ... and
f (x) = 0 for any x ∈ / {N, N + 1, N + 2, ...}. Using only the definition
of
R∞ an improper integral of the first type and theorem 19 we see that
a
f (x)dx is convergent and it is zero, in spite of that fact that f is
unbounded: f (N + i) = N + i → ∞, when i → ∞.
The following example is a standard example.
Example R47. Let a be a positive real number. For any α ∈ R we
∞
define I(α) = a x1α dx. Then this integral is convergent if and only if
α > 1. Indeed, if α = 1, I(1) = ln x |∞
a = ∞. Assume that α 6= 1. Then
Z x ¯x
1 t−α+1 ¯¯ x−α+1 a−α+1
F (x) = dt = = − .
a t
α −α + 1 ¯a −α + 1 −α + 1
Since lim x−α+1 is finite and 0 in this case, if and only if −α + 1 < 0,
x→∞
i.e. α > 1.
There is a great number of similarities between the improper in-
tegrals of the first type and the numerical series. First of all, for a
function
R x f : [a, ∞) → R+ , such that for any x > a the "proper" inte-
gral a f (t)dt exists, let us write formally (for a fixed natural number
N ≥ a):
Z ∞ Z N X∞
(2.2) f (x)dx = f (x)dx + cn ,
a a n=N
R n+1
where cn = n f (x)dx, n = N, N +1, ... . It is clear that the improper
integral is convergent if and only if the last numerical series on the
right is convergent (sketch a proof!). If f has positive and also negative
116 3. IMPROPER (GENERALIZED) INTEGRALS
values "up to ∞”, this statement is not always true. Indeed, let f (x) =
sin 2πx, x ∈ [0, ∞). Since
Z x ¯x
cos 2πt ¯¯ 1
F (x) = sin 2πtdt = − ¯ = [1 − cos 2x] ,
0 2π 0 2π
and since cos x hasR no limit at ∞, F (x) has no limit at ∞. Thus the
∞
improper integral 0 sin 2πxdx is not convergent.
R n+1
However, cn = n sin 2πxdx = 0 and the series on the right in
formula (2.2) is equal to zero!
We recall a nice theorem (theorem 21-the integral test, [Po]) which
is stated here in language of improper integrals.
Theorem 33. (series test) Let f : [a, ∞) → R+ , be a function f
defined on an interval [a, ∞), a ≥ 0, with nonnegative values. Assume
also that f is a continuous and a decreasing function. Let N be the
first natural number greater or equal to a andR let cn = f (n) for any
∞
n = N, N + 1, ... . Then the improper
P∞integral a f (x)dx is convergent
if and only if the numerical series n=N cn is convergent.
Since the proof of this result was done in [Po], theorem 21, we omit
it here. R∞
For instance, the integral a x1001 +1 dx, a ≥ 0, is convergent because
P
the numerical series ∞ 1
n=[a]+1 n100 +1 is convergent (apply the limit com-
parison test with the Riemann series for α = 100 > 1, etc.). To try
to compute directly a primitive for x1001 +1 is a crazy idea! And, if we
know that this integral is convergent but we cannot compute it ex-
actly, where is the practical gain? There is a big one! If we R Mknow that
1
this integral is convergent we can approximate it with a x100 +1 dx,
where M is a large natural number. But this last integral can be easily
approximately computed by using some specialized software.
If we try to use approximate methods for computing a divergent
improper integral is another crazy idea! This means to obtain different
numerical values which are randomly distributed to ∞. Thus these
values cannot approximate some unknown number!
This is way we need some tests in order to decide if an improper
integral is convergent or not. A basic test is obtained by applying the
Cauchy criterion for limits of functions (see theorem 31).
Theorem 34. (Cauchy criterion for improper integrals I)
Let f : [a, ∞) → R be a functionR ∞which is integrable on any finite
interval [a, x] for x ∈ [a, ∞). Then a f (x)dx is convergent if and only
if for any small ε > 0, there exists δ large enough and depending on ε,
2. IMPROPER INTEGRALS OF THE FIRST TYPE 117
¯R 00 ¯
¯ x ¯
such that if x0 , x00 ∈ [δ, ∞), then ¯ x0 f (x)dx¯ < ε. This is equivalent
R x00
to say that x0 f (x)dx → 0, whenever x0 , x00 → ∞.
R ∞ Example
arctan x
48. Let us study the convergence of the integral I =
1 x2
dx. Practically
¯R 00 we want
¯ to prove that the limit of the func-
0 00 ¯ x arctan x ¯
tion G(x , x ) = ¯ x0 x2
dx¯ is equal to zero, whenever x0 , x00 → ∞,
independently. Indeed, use the basic mean theorem (see theorem 22)
and find
¯Z 00 ¯ ¯Z 00 ¯
¯ x arctan x ¯ ¯ x 1 ¯ π ¯¯ 1 ¯
1 ¯¯
¯ ¯ ¯ ¯ ¯
¯ dx¯ = |arctan c| ¯ dx¯ ≤ − → 0,
¯ x0 x2 ¯ ¯ x0 x2 ¯ 2 ¯ x0 x00 ¯
when x0 , x00 → ∞.
This last test is very used to derive other more useful in practice
tests. In many cases the following R ∞ test reduces the study of the im-
proper integrals of the first type a Rf (x)dx to the case of the nonneg-
∞
ative Rfunction |f (x)| . If the integral a |f (x)| dx is convergent, we say
∞
that a f (x)dx is absolutely convergent.
Take now a small ε > 0 and let δ > 0 be large enough such that if
ε
x ∈ [δ, ∞) then |g(x)| < 4M . This is true because lim g(x) = 0. This δ
x→∞
is the searched for δ in Cauchy criterion:
¯Z 00 ¯
¯ x ¯ h ε ε i
¯ ¯ 00 0
¯ f (x)g(x)dx¯ ≤ 2M [|g(x )| + |g(x )|] < 2M + = ε,
¯ x0 ¯ 4M 4M
if x0 , x00 ∈ [δ, ∞). ¤
Example 49. R ∞(Dirichlet’s integral) If β is a real number, the fol-
lowing integral 0 sinxβx dx is called the Dirichlet’s integral. We can
assume that β > 0. Let us prove that this integral is convergent. First
of all let us remark that the function sinxβx defined initially only on
(0, ∞) can be extended to a continuous function
½ sin βx
, if x ∈ (0, ∞)
h(x) = x
β, if x = 0,
defined
R∞ on [0, ∞). Thus, we can substitute the initial integral with
0
h(x)dx. This last one is "identical" with the initial one because
does not matter what we put in x = 0, the value of the integral does
not change if that one is convergent. Since
Z ∞ Z 1 Z ∞
sin βx sin βx sin βx
dx = dx + dx,
0 x 0 x 1 x
R∞
it will be enough to prove that the last integral 1 sinxβx dx is convergent.
In order to prove this we apply the Dirichlet’s test R xby putting f (x) =
sin βx, which has the Newton’s primitive F (x) = 1 sin βtdt = cosβ β −
cos βx
β
bounded by β2 on [1, ∞), and g(x) = x1 , which is of class C 1 and
decreasing to 0 at ∞. Hence the Dirichlet’s integral is convergent for
any fixed β ∈ R.
Example
R ∞ 50. (Fresnel’s integrals)
R∞ The following integrals
2 2
I = 0 sin x dx and J = 0 cos x dx are called the Fresnel’s inte-
grals. Let us prove that I is convergent (in fact both are convergent!).
For this, let us write
Z ∞ Z 1 Z ∞
2 2
sin x dx = sin x dx + sin x2 dx.
0 0 1
R∞ 2
R x to2 prove that 1 sin x dx is convergent.√ Let us denote
It is enough
F (x) = 1 sin t dt and let us change the variable t = u. So
Z x Z 2
2 1 x 1
sin t dt = √ sin udu.
1 2 1 u
3. IMPROPER INTEGRALS OF THE SECOND TYPE 123
2
Since any real number z ∈ [1, ∞) can be written
R ∞ 1 as z = x for an x ∈
[1, ∞), it is enough to see that the integral 1 √u sin udu is convergent.
Here we apply again the Dirichlet’s test for f (u) = sin u and g(u) = √1u .
So the first Fresnel’s integral is convergent. We leave as an exercise
for the reader to prove (in the same way!) that the second Fresnel’s
integral is also convergent.
If f is integrable on any subinterval [y, b], where y ∈ (a, b], then again
f is integrable on the entire interval [a, b] and
Z b Z b
(3.2) f (x)dx = lim f (t)dt
a y→a,y>a y
Proof. Since the proof for the second part of the statement is very
similar to the proof of the first part, we shall prove only this last one.
To prove the integrability of f on [a, b] we shall use Darboux Criterion
(theorem 14). Since f is bounded on [a, b] one can choose M > 0 such
that |f (x)| ≤ M on [a, b]. Let ε > 0 be a small positive real number
and let δ 1 = ε/2M. Let δ > 0 such that δ ≤ δ 1 and for any partition
∆ = {a = x0 , x1 , ..., xn−1 = b − δ 1 } of the interval [a, b − δ 1 ] with
k∆k < δ, one has:
X
n−1
S∆ − s∆ = (Mi − mi ) (xi − xi−1 ) < ε/2.
i=1
of the interval [a, b]. If we keep the above δ fixed, any partition ∆00 of
[a, b], k∆00 k < δ, can be represented as a partition ∆0 like in formula
(3.3), with a δ 01 ≤ δ1 , instead of δ 1 . Thus
S∆0 − s∆0 ≤ S∆ − s∆ + (Mn − mn )δ 1 ≤ ε/2 + M · ε/2M = ε.
Applying Darboux Criterion we get that f is integrable and
Z b
f (x)dx = lim
0
S∆0 = lim S∆ =
a k∆ k→0 k∆k→0
Z b−δ 1 Z x
= lim f (x)dx = lim f (t)dt.
δ 1 →0 a x→b,x<b a
¤
For instance, applying this last result we find that the noncontinu-
ous function f : [0, 1] → R,
½
x, if x ∈ [0, 1)
f (x) = ,
2, if x = 1
is integrable and
Z 1 Z x
x2 1
f (x)dx = lim tdt = lim = .
0 x→1,x<1 0 x→1,x<1 2 2
We say that a function f : I → R, defined on an interval I with
real values is unbounded at a limit point b of I if there exists at least
one sequence {xn } with xn ∈ I, xn → b such that the sequence {f (xn )}
is not bounded. For instance, f (x) = 1/x , x ∈ (0, 1] is not bounded
at 0. We also say that 0 is a singular point for f. We can also write
lim x1 = ∞, thus f (x) = 1/x is not bounded at 0. In our definition
x→0,x>0
we can take xn = 1/n.
Definition 11. Let [a, b] be a bounded and closed interval in R
and let f : [a, b) → R such that f is integrable on any subinterval [a, x],
x ∈ [a, b) and f is unbounded R x at b. We say that f is integrable on [a, b] if
the following limit lim a f (t)dt exists, is finite and in this case we
x→b,x<b
Rb
denote it by a f (t)dt. This last integral is called an improper integral
R b the second type with singularity at b. We also say that the integral
of
a
f (t)dt is convergent.
Exercise 7. Let f : (a, b] → R be an integrable function on any
interval [x, b], x ∈ (a, b] such that f is unbounded at a. Define the
Rb
convergence of a f (t)dt in this last case.
3. IMPROPER INTEGRALS OF THE SECOND TYPE 125
R1
For instance, f (x) = 1/x, x ∈ (0, 1] is unbounded at 0 and 0 x1 dx
is divergent because
Z 1
1
lim dt = lim [ln 1 − ln x] = ∞.
x→0,x>0 x t x→0,x>0
R1 1
But 0 √x dx is convergent because
Z 1 ¯1 h i
1 1¯ 1
lim 2 ¯
√ dt = lim 2t ¯ = lim 2 − 2x = 2. 2
x→0,x>0 x t x→0,x>0
x
x→0,x>0
If this last limit exists, in practice one can work with it, but,...very
carefully, because Rits value is only a convention. R0
1
For instance, −1 x1 dx is divergent because neither −1 x1 dx, nor
R1 1
0 x
dx is convergent.
126 3. IMPROPER (GENERALIZED) INTEGRALS
R1
However, the principal value of −1 x1 dx is 0. Indeed,
Z 1 ∙Z −ε Z 1 ¸
1 1 1
dx = lim dx + dx =
−1 x ε→0,ε>0 −1 x ε x
= lim [ln ε − ln 1 + ln 1 − ln ε] = 0.
ε→0,ε>0
To use in practice the definition 11 is usually R xa very difficult task
because we must compute integrals of the type a f (t)dt. In practical
problems
Rb we want to know if our improper integral of the second type
f (x)dx with singularity at b is convergent or not. If it is convergent,
a R b−ε
then we can approximate it with a f (x)dx for a very small chosen
positive number ε.RFor instance, ε = 10−6 . Now, to compute this last
b−ε
Riemann integral a f (x)dx one can use any approximate formula
just studied in Chapter 2. To R b decide the convergence of an improper
integral of the second type a f (x)dx with singularity at b, we need
some "tests".
We begin with a standard example.
Rb 1
Example 51. Let I(α) = a (b−x) α dx be an improper integral of the
But Z x
1 (b − x)−α+1 (b − a)−α+1
α
dt = − + ,
a (b − t) −α + 1 −α + 1
if α 6= 1, and Z x
1 b−a
α
dt = ln ,
a (b − t) b−x
if α = 1. It is easy to see that the limit in (3.4) exists if and only if
−α+1
α < 1. In this last case I(α) = (b−a) −α+1
.
Rx
If F (x) = a f (t)dt, x ∈ [a, b) is the area function of Newton for
Rb
function f which appears in the last definition, then a f (t)dt is con-
vergent if and only if the function F : [a, b) → R has a (unique) finite
limit at the point b. Thus we can apply the general Cauchy Criterion
(see theorem 31, a)) for the function F and obtain:
Theorem 41. (Cauchy criterion for improper integrals II) Let f :
[a, b) → R be a function which is integrable on any finite interval [a, x]
Rb
for x ∈ [a, b). Then a f (x)dx is convergent if and only if for any small
3. IMPROPER INTEGRALS OF THE SECOND TYPE 127
For instance,
R 1 sin( √
let us use this last result to prove that the integral
4 x)
I = 0 √x dx, with the unique singularity x = 0, is convergent. We
shall prove that it is absolutely convergent. Indeed,
Z x00 √ ¯Z 00 ¯
|sin( 4 x)| ¯ x 1 ¯ ¯ √ √ ¯¯
¯ ¯ ¯
√ dx ≤ ¯ √ dx¯ = ¯2 x00 − 2 x0 ¯ → 0,
x0 x ¯ x0 x ¯
convergent. Since
Z b Z b−δ Z b
(3.6) f (x)dx = f (x)dx + f (x)dx,
a a b−δ
Rb
one has that a f (x)dx is convergent.
ii) If L = 0, in formula (3.5) we use only the last inequality to see
Rb Rb Rb
that if a g(x)dx is convergent, then ε a g(x)dx and ε b−δ g(x)dx are
Rb
also convergent and b−δ f (x)dx is convergent. From formula (3.6) we
Rb
get that a f (x)dx itself is convergent.
iii) If L = ∞, then for any fixed M > 0, there exists a small
number δ > 0 such that fg(x)
(x)
> M if x ∈ [b − δ, b). Since f (x) > Mg(x),
Rb Rb
x ∈ [b−δ, b) and since a g(x)dx and M b−δ g(x)dx are divergent, then
Rb
b−δ
f (x)dx is also divergent. Now we apply again formula (3.6) to see
Rb
that a f (x)dx is divergent. ¤
For instance, since
√ 1
x(x+1)
lim =1
x→0,x>0 √1
x
R2
and since √1 dx
is convergent (α = 1/2 < 1 in example 51), this last
0 x
R2 1
theorem says that 0 √x(x+1) dx is convergent.
1
If in theorem 44 we take g(x) = (b−x)q , we get a very useful test.
if x0 , x00 ∈ [b − δ, b). ¤
sin x
[0, ∞) with function f (x) = √x+sin x
. It is easy to see that this function
could be unbounded only at x = 0. But
3
sin x x − x3! + ...
lim √ = lim √ 3 = 0,
x→0,x>0 x + sin x x→0,x>0 x + x − x + ...
3!
so that our function is bounded on [0, ∞) and the integral is an improper
integral of the first type, i.e. it has a singularity only at ∞. Let us write
I = I1 + I2 ,
R2 sin x
R ∞ sin x
where I1 = 0 √x+sin x
dx and I2 = 2 √x+sin x
dx. In order to prove
that I is convergent (divergent) it is enough to prove that I2 is conver-
gent (divergent) because I1 is a proper integral, i.e. an usual Riemann
integral on [0, 2]. And any proper integral is convergent! Let us write
now
sin x sin x sin2 x
√ = √ −√ √ .
x + sin x x x( x + sin x)
R∞ x
Applying Dirichlet’s test (theorem 39) to the integral 2 sin √ dx we get
x
that this one is convergent. Thus theR nature of the integral I2 is the
∞ 2x
same like the nature of the integral 2 √x(√sinx+sin x)
dx. But this time
sin2 x
the integrand function √ √
x( x+sin x)
is nonnegative and we can compare
2
it with an easier nonnegative function g(x) = √x(sin√x+1) x
≤ f (x) =
2
sin x
R ∞ 2
√ √
x( x+sin x)
x
. It is enough to prove that 2 √x(sin√x+1) dx is divergent to
R∞ sin2x
conclude that 2 √x(√x+sin x) dx is divergent and so I2 would be diver-
gent. Since the integral
Z Z
1 ∞ 1 − 2 sin2 x 1 ∞ cos 2x
− √ √ dx = − √ √ dx
2 2 x( x + 1) 2 2 x( x + 1)
is convergent
R (apply again Dirichlet’s test) and since the integral
1 ∞ √ √1
2 2 x( x+1)
dx is divergent (apply test p for p = 1), from
Z ∞ Z Z
sin2 x 1 ∞ 1 − 2 sin2 x 1 ∞ 1
√ √ dx = − √ √ dx+ √ √ dx,
2 x( x + 1) 2 2 x( x + 1) 2 2 x( x + 1)
R∞ 2x
we conclude that the integral 2 √x(sin√x+1) dx is divergent. Thus I2 and
I are also divergent. This information is very important because it will
be a crazy idea to approximate with something a divergent integral. You
can say "maybe the given integral has something wrong in√its form if
somebody
√ insists to compute it!".R For instance, if instead of x one put
4 3 ∞ sin x
x the new obtained integral 0 √ 4 3
x +sin x
dx were convergent (prove
this following the same steps as above!).
134 3. IMPROPER (GENERALIZED) INTEGRALS
(1.3)
Z a(t0 ) Z b(t0 ) Z b(tn ) Z b(t0 )
f (x, tn )dx+ f (x, tn )dx+ f (x, tn )dx− f (x, t0 )dx.
a(tn ) a(t0 ) b(t0 ) a(t0 )
R a(t ) R b(t )
To evaluate the integrals a(tn0) f (x, tn )dx and b(t0n) f (x, tn )dx we apply
the mean formula (see theorem 22) and find:
Z a(t0 )
(1.4) f (x, tn )dx = f (cn , tn ) [a(t0 ) − a(tn )]
a(tn )
and
Z b(tn )
(1.5) f (x, tn )dx = f (dn , tn )[b(tn ) − b(t0 )],
b(t0 )
where cn is between a(tn ) and a(t0 ) and dn is between b(tn ) and b(t0 ).
Since functions a(t) and b(t) are continuous and tn → t0 , we see that
the sequence {cn } converges to a(t0 ) and the sequence {dn } tends to
b(t0 ). Since f is a continuous function as a function of two variables,
we see that f (cn , tn ) → f (a(t0 ), t0 ) and f (dn , tn ) → f (b(t0 ), t0 ). But
a(t0 ) − a(tn ) → 0 and b(tn ) − b(t0 ) → 0, so that
Z a(t0 )
f (x, tn )dx → f (a(t0 ), t0 ) · 0 = 0
a(tn )
and
Z b(tn )
f (x, tn )dx → f (b(t0 ), t0 ) · 0 = 0.
b(t0 )
for any t ∈ J.
Proof. Let us fix t0 in J and let us evaluate the ratio
R b0
I(t) − I(t0 ) [f (x, t) − f (x, t0 )] dx
= a0 .
t − t0 t − t0
For any fixed x in [a0 , b0 ] we define gx : [t, t0 ]± → R by gx (ζ) = f (x, ζ).
Here [t, t0 ]± = [t, t0 ] if t ≤ t0 and [t, t0 ]± = [t0 , t] if t0 ≤ t. Let us write
Lagrange’s formula for the function gx on the interval [t, t0 ]± (i.e. we
1. PROPER INTEGRALS WITH PARAMETERS 141
Since ∂f
∂t
(x, cx,t ) is a new continuous function h(x, t) of two variables,
we can apply the continuity theorem 48 to the proper integral
Z b0 Z b0
∂f
(x, cx,t )dx = h(x, t)dx
a0 ∂t a0
with parameter t and obtain
Z b0 Z b0
I(t) − I(t0 ) ∂f
lim = lim h(x, t)dx = (x, t0 )dx,
t→t0 t − t0 a0 t→t0 a0 ∂t
for any t ∈ J.
Proof. Let us fix a point t0 ∈ J and let us evaluate the ratio
R b0 Rb R a(t0 )
I(t) − I(t0 ) a(t)
f (x, t)dx − a(t0 0 ) f (x, t0 )dx a(t)
[f (x, t)] dx
= = +
t − t0 t − t0 t − t0
R b0
a(t )
[f (x, t) − f (x, t0 )] dx
+ 0 .
t − t0
The limit of R b0
a(t0 )
[f (x, t) − f (x, t0 )] dx
t − t0
when t → t0 is exactly
Z b0
∂f
(x, t0 )dx
a(t0 ) ∂t
(see the proof of theorem 49). Thus, it remains to prove that the
U a(t0 )
[f (x,t)]dx
limit of a(t) t−t0 when t → t0 is equal to −a0 (t0 )f (a(t0 ), t0 ). To
do this, let us apply mean formula of theorem (22) to the integral
R a(t0 )
a(t)
[f (x, t)] dx :
Z a(t0 )
[f (x, t)] dx = f (dt , t) [a(t0 ) − a(t)] ,
a(t)
where dt ∈ [a(t), a(t0 )]± . Since a(t) is also continuous, a(t) → a(t0 )
and so dt → a(t0 ), whenever t → t0 . Thus,
R a(t0 )
a(t)
[f (x, t)] dx a(t0 ) − a(t)
lim = f (a(t0 ), t0 )· lim = −a0 (t0 )f (a(t0 ), t0 ).
t→t0 t − t0 t→t0 t − t0
Finally we put together both limits computed above and obtain:
R a(t0 ) R b0
I(t) − I(t0 ) a(t)
[f (x, t)] dx a(t0 )
[f (x, t) − f (x, t0 )] dx
lim = lim +lim =
t→t0 t − t0 t→t0 t − t0 t→t0 t − t0
Z b0
0 ∂f
−a (t0 )f (a(t0 ), t0 ) + (x, t0 )dx,
a(t0 ) ∂t
i.e. we just obtained formula (1.8) for t = t0 . ¤
Let us apply this last formula to compute the following limit:
R 1 x2
e dx
limπ R sin
0
t
.
t→ 2 ex2 dx
cos t
1. PROPER INTEGRALS WITH PARAMETERS 143
Proof. First of all let us see that this formula generalizes both
formulas (1.7) and (1.8). We shall see that these last formulas are
enough in order to deduce the general Leibniz formula (1.9). Let us fix
a t0 ∈ J and let us write
Z b(t) Z a(t0 ) Z b(t0 ) Z b(t)
I(t) = f (x, t)dx = f (x, t)dx+ f (x, t)dx+ f (x, t)dx.
a(t) a(t) a(t0 ) b(t0 )
Proof. 1) First of all let us explain what formula (1.12) says. Let
Rd
us denote J(x) = c f (x, t)dt, a proper integral with a parameter x and
Rb
let denote by L(t) = a f (x, t)dx, a proper integral with a parameter
Rb Rd
t. Formula (1.12) says that a J(x)dx = c L(t)dt, i.e. we may change
the order of integration.
R u ³R d ´
For proving Fubini’s formula we put F (u) = a c f (x, t)dt dx
R d ¡R u ¢
and G(u) = c a f (x, t)dx dt for any u ∈ [a, b]. If we succeed to
prove that F (u) = G(u) for any u ∈ [a, b], by making u = b, we would
obtain formula (1.12). Let us carefully use Leibniz formula (1.9) to
compute F 0 (u) and G0 (u).
Z d Z d
0 0
F (u) = f (u, t)dt, G (u) = f (u, t)dt.
c c
Thus, F 0 (u) = G0 (u) for any u ∈ [a, b]. Hence F (u) = G(u) + C, where
C is a constant w.r.t. u. Let us put u = a and find F (a) = G(a) + C
and, since F (a) = G(a) = 0, we find that C = 0, i.e. F (u) = G(u) and
the first proof is completed. Here is another proof.
2) Let us consider two arbitrary divisions
∆x : a = x0 < x1 < ... < xn = b
and
∆y : c = t0 < t1 < ... < tm = d
of the intervals [a, b] and [c, d] respectively. Let us denote by M1 =
Rb Rd
a
J(x)dx and by M 2 = c
L(t)dx. We can approximate very well the
number M1 by Riemann sums of the form
X
n
(1.13) J(ξ i )(xi − xi−1 ),
i=1
Since the sum and the product are commutative we see that the above
double sums are equal. Hence the numbers M1 and M2 can be well
approximate by the one and the same set of numbers of the form
Xn Xm
f (ξ i , η j )(tj − tj−1 )(xi − xi−1 ).
i=1 j=1
is called a Riemann double sum for the function f. If all this double
sums becomes closer and closer to a number I, we say that I is the
ZZ
double integral of f on the rectangle D. Write I = f (x, t)dxdt.
D
2. IMPROPER INTEGRALS WITH PARAMETERS 147
p with p = 2), so the Cauchy criterion condition works, one has that
for¯ any small ε¯ > 0 there exists M > 0 such that if x0 , x00 > M we get
R 00 1
π ¯ x ¯
2 ¯ x0 u(u+1)
du¯ < ε. Thus, since
¯Z 00 ¯ ¯Z 00 ¯
¯ x arctan tu ¯ Z x00 ¯¯ arctan tu ¯¯ π ¯ x 1 ¯
¯ ¯ ¯ ¯ du ≤ ¯¯ ¯
¯ du¯ ≤ ¯ u(u + 1) ¯ du¯<ε
¯ x0 u(u + 1) ¯ x0 2 ¯ x0 u(u + 1) ¯
¯R 00 ¯
¯ x ¯
we obtain that ¯ x0 arctan tu
u(u+1)
du ¯ < ε for any x0 , x00 > M. Hence
Z ∞ Z 1 Z ∞
arctan tu arctan tu arctan tu
g(t) = du = du + du
0 u(u + 1) 0 u(u + 1) 1 u(u + 1)
is the uniform limit of f (x, t) relative to the parameter t.
Theorem 54. (continuity theorem) Let f (x, t), f : K × J → R,
where K is a real interval and J is the domain of parameters t =
(t1 , t2 , ..., tn ), J ⊂ Rn . Assume that f is continuous with respect to the
variable vector t and that f has a uniform limit g(t) at a limit point
b of K. Then g(t) = limf (x, t) is a continuous function of its vector
x→b
variable t. Thus, for any fixed t0 ∈ J,
h i
(2.1) lim limf (x, t) = limf (x, t0 ).
t→t0 x→b x→b
for any x ∈ Vb and for any t ∈ J. Now let us fix such an x0 ∈ Vb and
write again the inequality (2.2) for x = x0 :
(2.4)
|g(t) − g(t0 )| ≤ |g(t) − f (x0 , t)|+|f (x0 , t) − f (x0 , t0 )|+|f (x0 , t0 ) − g(t0 )| .
Since function t → f (x0 , t) is continuous at t0 , there is a small real
number δ > 0 such that if kt − t0 k < δ, we get
ε
(2.5) |f (x0 , t) − f (x0 , t0 )| < .
3
Let us use the three inequalities (2.3) and (2.5) in the evaluation (2.4)
to finally obtain:
|g(t) − g(t0 )| < ε
for any t ∈ J with kt − t0 k < δ. But this exactly means the continuity
of the function g at t0 . ¤
R ∞ arctan tu
Example 56. Let us take the function g(t) = 0 u(u+1) du (this
is an improper integral with a parameter R x t, seetuthe definition bellow)
obtained as a uniform limit of f (x, t) = 0 arctan u(u+1)
du when x → ∞ and
let us compute Z ∞
arctan tu
limg(t) = lim du.
t→0 t→0 0 u(u + 1)
Since g is continuous (see the last theorem above), we can write 0 =
g(0) = limg(t). Thus this limit is equal to 0. We simply put x = 0
t→0
under the integral sign. But, we must be very careful and not to do this
before you did not prove that the limit is uniform!
With the supplementary condition that f (x, t) has a finite limit at
a limit point t0 of J, one can extend formula (2.1) for this more general
case. For instance, if a > 0, then
Z ∞ Z ∞
arctan tu arctan ∞ π a
lim du = du = − ln .
t→∞ a u(u + 1) a u(u + 1) 2 a+1
(What happens when a → 0?).
The above theorem is sometimes used to decide if a limit is uniform
or not. For instance, take f (x, t) = tx , where x ∈ R and t ∈ (0, 1]. If
g(t) = lim f (x, t), it it easy to see that
x→∞
½
0, if t < 1
g(t) = .
1, if t = 1
Since this function in not a continuous function, the above limit cannot
be uniform with respect to the parameter t. This is also a counterex-
ample to the fact that the hypothesis on limit to be a uniform one is
necessary.
2. IMPROPER INTEGRALS WITH PARAMETERS 151
Z b
0 ∂f
(2.8) I (t) = (x, t)dx − a0 (t)f (a(t), t),
a(t) ∂t
for any t ∈ J (Leibniz formula).
154 4. INTEGRALS WITH PARAMETERS
Rb
e) (Weierstrass test) Let I(t) = a f (x, t)dx, a ∈ R, be an improper
integral with a parameter t and with singularity b. If there exists a
function s(x), s : K → R such that |f (x, t)| ≤ s(x) for any x ∈ K,
Rb
t ∈ J and if the improper integral a s(x)dx is convergent, then I(t) is
uniformly convergent.
Proof. Let us prove for instance Leibniz formula d).
Rx Rb
Let F (x, t) = a(t) f (u, t)du. Since the integral I(t) = a(t) f (x, t)dx
is convergent, the function F (x, t) has as a limit at b, exactly the func-
Rb
tion I(t). Since the integral H(t) = a(t) ∂f ∂t
(x, t)dx is uniformly conver-
R x ∂f
gent, the function L(x, t) = a(t) ∂t (u, t)du has as a uniform limit the
function H(t). Applying theorem 56 we get that
∙Z x ¸
0 ∂F ∂f 0
I (t) = lim (x, t) = lim (u, t)du − a (t)f (a(t), t) =
x→b ∂t x→b a(t) ∂t
Z b
∂f
= (x, t)du − a0 (t)f (a(t), t).
a(t) ∂t
Rx
Here we just applied Leibniz formula for the proper integral a(t) f (u, t)du
with parameter t.
Let us prove now Weierstrass test e). In order to prove that the
Rb
improper integral I(t) = a f (x, t)dx with parameter t is uniformly
convergent, we use Cauchy criterion a) of the same theorem. Let ε > 0
be a small real number and let Vb¯ be a neighborhood ¯ of b in K such
0 00 ¯R x00 ¯
that for any x , x ∈ Vb one has ¯ x0 s(x)dx¯ < ε (here we used the
Rb
fact that the improper integral a s(x)dx is convergent). Since
¯Z 00 ¯ Z 00 Z x00
¯ x ¯ x
¯ ¯
¯ f (x, t)dx¯ ≤ |f (x, t)| dx ≤ s(x)dx < ε,
¯ x0 ¯ x0 x0
¯R 00 ¯
¯ x ¯
we finally derive that ¯ x0 f (x, t)dx¯ < ε for any x0 , x00 ∈ Vb . Applying
again the Cauchy criterion a), we get that the integral I(t) is uniformly
convergent. ¤
Let us use Leibniz formula d) of this last theorem (forR improper in-
∞
tegrals with a parameter) to compute the integral I(t) = 0 arctan tx
x(1+x2 )
dx,
t > 0. We can easily see that this integral is uniformly convergent (for
applying Leibniz formula this is not necessarily!). To see this we write
¯ ¯
¯ arctan tx ¯ π 1
¯ ¯
¯ x(1 + x2 ) ¯ ≤ 2 x(1 + x2 ) .
2. IMPROPER INTEGRALS WITH PARAMETERS 155
R∞ 1
Since the improper integral π2 1 x(1+x 2 ) dx is convergent (apply p-
test with p = 3), Weierstrass test e) in the above theorem says that
R ∞ arctan tx
1 x(1+x2 )
dx is uniformly convergent. Since lim arctan tx
2 = t, the in-
x→0 x(1+x )
R 1 arctan tx
tegral 0 x(1+x2 ) dx is a proper integral. It is easy to see that a sum
between a proper integral and an uniformly convergent improper in-
tegral is also an uniformly convergent improper integral (prove this!).
Thus
Z ∞ Z 1 Z ∞
arctan tx arctan tx arctan tx
I(t) = 2
dx = 2
dx + dx
0 x(1 + x ) 0 x(1 + x ) 1 x(1 + x2 )
is an uniformly convergent integral. We also need that the integral
Z ∞ ∙ ¸ Z ∞
∂ arctan tx 1
H(t) = 2
dx = dx
0 ∂t x(1 + x ) 0 (1 + x )(1 + t2 x2 )
2
1 1
≤
(1 + x2 )(1 + t2 x2 ) 1 + x2
R∞ 1
and since 1 (1+x 2 ) dx is convergent (test p for p = 2), applying again
R∞ 1
Weierstrass test e) we get that 1 (1+x2 )(1+t 2 x2 ) dx is uniformly conver-
But
1 1 1 t2 1
2 2 2
= 2
· 2
− ·
(1 + x )(1 + t x ) 1−t 1+x 1 − t 1 + t2 x2
2
1 t π 1
= 2
arctan x|∞0 − 2
arctan tx |∞ 0 = .
1−t 1−t 21+t
156 4. INTEGRALS WITH PARAMETERS
for any x0 , x00 R> M. Thus, the Weierstrass test e) tells us that the
∞
integral I(t) = 1 sinx2tx dx is uniformly convergent.
Now we give some applications to the effective computation of some
classes of integrals with parameters.
Example 57. Let us compute the value of the following integral
with two parameters α and β:
Z ∞ −αx2 2
e − e−βx
(2.9) I(α, β) = dx, α, β > 0.
0 x
Since
2 2 2 2
e−αx − e−βx 0 −2αxe−αx − 2βxe−βx
lim = = lim = 0,
x→0 x 0 x→0 1
for any fixed α and β, the integral has only b = ∞ as a singularity.
Considering successively I(α, β) as an integral with a parameter α and
then, with a parameter β respectively, we can apply Leibniz formula
separately for α and β respectively. Let us prove the simple convergence
of I(α, β). Since
−αx2 −βx2
2e 2e
lim x = 0, lim x = 0,
x→∞ x x→∞ x
if α, β > 0, the p-test with p = 2 gives us the simple convergence of
I(α, β). We need now the uniform convergence of
Z ∞ " 2 2
# Z ∞
∂I ∂ e−αx − e−βx 2
= dx = − xe−αx dx.
∂α 0 ∂α x 0
2. IMPROPER INTEGRALS WITH PARAMETERS 157
For any fixed α1 > 0, this last integral is uniformly convergent on the
interval J = [α1 , ∞). Indeed,
¯ ¯
¯ −αx2 ¯ 2 2
¯xe ¯ = xe−αx ≤ xe−α1 x
R∞ 2
and 0 xe−α1 x dx is convergent (apply test p for p = 2), thus applying
again
R ∞ Weierstrass test e) of theorem 57 we obtain that the integral
−αx2
− 0 xe dx is uniformly convergent w.r.t. α. In the same way we
can prove that the integral
Z ∞ " 2 2
# Z ∞
∂I ∂ e−αx − e−βx 2
= dx = xe−βx dx
∂β 0 ∂β x 0
type [δ, ∞) and not on the entire interval (0, ∞). We shall prove now
that the integral
Z ∞
cos ax − cos bx b
I(a, b) = dx = ln
0 x a
is not uniformly convergent w.r.t. a and b (simultaneously), where
a, b ∈ (0, ∞). Suppose it is uniformly convergent. Applying theorem 57
1
b) we get for a = 2n , b = n1 and n → ∞ that 0 = ln 2, a contradiction!
Thus our integral cannot be uniformly convergent on (0, ∞). It is an
open problem for me if this integral is uniformly convergent on any
interval of the type [δ, ∞), where δ > 0. But on intervals of the type
[δ, M], 0 < δ < M? These informations could not help us to compute
the integral by Leibniz formula because the integral
Z ∞ ∙ ¸ Z ∞
∂ cos ax − cos bx
dx = − sin axdx
0 ∂a x 0
is not convergent!
Example 58. (Dirichlet’s integral) The following integral with a
parameter β,
Z ∞
sin βx
I(β) = dx
0 x
is very useful in different branches of science and technique. It was
considered for the first time by the great German mathematician Jo-
hann Peter Gustave Lejeune Dirichlet in the XIX-th century. To prove
its existence, for any fixed β, he invented the Dirichlet test (see the-
orem 39). In example 49 we just proved its simple convergence. In
order to compute it we need it to be uniformly convergent. Even we
had succeeded to prove such thing, Leibniz formula leads us to compute
Z ∞ ∙ ¸ Z ∞
∂ sin βx
dx = cos βxdx,
0 ∂β x 0
R ∞ Example
−x2
59. (Euler-Poisson integral) The famous integral J =
0
e dx, which appears in Probability and Statistics whenever we
work with the Gauss-Laplace distribution and in many other places of
Mathematics, Mechanics, etc. is called the Euler-Poisson integral. To
compute it (the p-test for p = 2 assure us on the convergence!) let us
introduce a parameter u and a new variable t.
R ∞Let−uus2 t2make a change of
variables x ↔ t, x = ut in J. We get J = u 0 e dt. Thus
Z ∞
2 2 2 2
e−u J = ue−u e−u t dt.
0
Let us integrate this last equality with respect to u and then let us change
the order of integration (see Fubini’s theorem for improper integrals,
theorem 57 c); verify the uniform convergence!):
Z ∞ Z ∞ Z ∞
−u2 2 −u2 2 2
J e du = J = ue du · e−u t dt =
0 0 0
Z ∞ µZ ∞ ¶ Z ∞ ¯∞
−u2 (t2 +1) 1 −(1+t2 )u2 ¯
= e u du dt = − e ¯ dt =
0 0 0 2(t2 + 1) 0
Z ∞
1 1 1 π
= 2
dt = arctan t |∞
0 = .
2 0 1+t 2 4
164 4. INTEGRALS WITH PARAMETERS
Hence
Z ∞ √
−x2 π
(2.21) J= e dx =
0 2
√
If we put x = tu, u being a new variable ant t > 0 a parameter, we
get
Z ∞ √
−tu2 1 π
(2.22) e du = √ .
0 t 2
R ∞ Example 60. (Fresnel’s
R∞ integrals) The following integrals I1 =
0
sin x dx and I2 = 0 cos x2 dx are called Fresnel’s integrals. Let
2
fractions. Since
¡ ¢2 √ √
t4 + 1 = t2 + 1 − 2t2 = (t2 + 1 − t 2)(t2 + 1 + t 2),
we get
1 At + B Ct + D
4
= √ + √ ,
1+t 2
t +1−t 2 t +1+t 2 2
Thus,
Z √ h
∞
1 2 π π π πi π
4
dt = 0 + + + − = √ .
0 1+t 4 2 4 2 4 2 2
p
π
Finally, I1 = √1π · 2√ = 12 π2 . We leave as an exercise for the reader
R ∞2 p
to prove that I2 = 0 cos x2 dx is also 12 π2 .
Remark 19. Sometimes we meet the following situation. Let K be
an interval [a, b] in R ∪ {±∞} and let J = [t0 , T ] be an interval in R ∪
{∞} such that t0 ∈ R. Let f (x, t) be a function of two variables defined
Rb
on K ×J such that it is of class C 1 on K ×J. Let I(t) = a f (x, t)dx be
an improper integral with a parameter t with a or b, or both of them as
its unique singular points.R Assume that for any δ > 0, δ < T − t0 , the
b
integrals I(t) and L(t) = a ∂f (x, t)dx are uniformly convergent w.r.t.
∂t Rb
t on [t0 + δ, T ]. We also assume that I(t0 ) = a f (x, t0 )dx is convergent
Rb
and that L(t) = a ∂f (x, t)dx is bounded in any neighborhood [t0 , t0 +
∂t Rb
δ] of t0 . Then the function t → I(t) = a f (x, t)dx is well defined
on [t0 , ∞) and it is continuous on this interval. Indeed, since I(t)
is uniformly convergent on any interval of the type [t0 + δ, T ] it is
continuous on (t0 , T ] (see theorem 57 b)). Now we prove that we can
"extend by continuity" this continuous function I(t), t ∈ (t0 , T ] to the
Rb
entire interval [t0 , T ] by putting I(t0 ) = a f (x, t0 )dx. Thus we want
to prove the continuity of I at t0 in [t0 , T ]. Take a sequence {tn }n≥1 ,
tn ∈ (t0 , T ] and tn → t0 . Let us evaluate the difference I(tn ) − I(t0 ),
when n → ∞. For any fixed x ∈ (a, b), let us apply Lagrange formula
for the function t → f (x, t) on the interval [t0 , tn ] :
∂f
f (x, tn ) − f (x, t0 ) = (x, cx,tn )(tn − t0 ),
∂t
where cx,tn ∈ [t0 , tn ]. Thus,
Z b Z b
∂f
I(tn ) − I(t0 ) = [f (x, tn ) − f (x, t0 )] dx = (tn − t0 ) (x, cx,tn )dx
a a ∂t
Rb
is convergent to 0 when n → ∞ because a ∂f ∂t
(x, cx,tn )dx is bounded
around the point t0 . Hence I(t) is also continuous at t = t0 .
R∞
For instance, let I(t) = 0 sinx x e−tx dx, t ∈ [0, ∞). It is not difficult
to prove that all conditions which appears in remark 19 are satisfied
for t0 =R0. Thus we can say that I is also continuous at t = 0. Its value
∞
I(0) = 0 sinx x dx = π2 (see Dirichlet’s integral 2.19).
Example 61. Let us study and compute the following integral
3. EULER’S FUNCTIONS GAMMA AND BETA 167
R1 xα −xβ
I(α, β) = 0 ln x
α, β > −1. For this let us formally write
dx,
Z 1 α Z 1 β
x −1 x −1
I(α, β) = dx − dx.
0 ln x 0 ln x
It is sufficient to prove that any of this last two integrals are proper
integrals (do it!). Thus I(α, β) is a proper integral and
Z 1 α Z 1 µZ α ¶ Z α µZ 1 ¶
x − xβ t t
dx = x dt dx = x dx dt =
0 ln x 0 β β 0
Z ¯1 Z α
α
xt+1 ¯¯ 1 α+1
= ¯ dt = dt = ln .
β t+1 0 β t+1 β+1
Here we just applied Fubini theorem 52, because all integrals which
appear can be considered proper integrals.
(k) (k)
Γ(k) (x) = Γ1 (x) + Γ2 (x),
where Z 1
(k)
Γ1 (x) = tx−1 (ln t)k e−t dt
0
and Z ∞
(k)
Γ2 (x) = tx−1 (ln t)k e−t dt.
1
If we could prove that these two last improper integrals with parameter
x are uniformly convergent on [δ, M], their sum Γ(k) (x) would also be
(k)
uniformly convergent on the same interval. Let us start with Γ1 (x). It
is an improper integral of the second type with a singular point t = 0.
Since
µ ¶k
¯ x−1 ¯ 1
¯t (ln t) e ¯ ≤ t |ln t| e = t
k −t δ−1 k −t δ−1
ln e−t ,
t
applying Weierstrass test e) of theorem 57, it will be enough to prove
R1 ¡ ¢k
the convergence of the integral 0 tδ−1 ln 1t e−t dt. We can assume
that δ is sufficiently small (we prove the uniform convergence on a
larger interval if we diminish δ!), for instance that 0 < δ < 1. Let us fix
now a number q such that 0 < q < 1, q + δ − 1 > 0 and another small
number ε with 0 < ε < q−1+δ k
. We did this such that the following limit
to be 0!
µ ¶k ¡ 1 ¢k
1 ln
(3.3) lim tq tδ−1 ln e−t = lim tq tδ−1 t−kε ¡ ¢tkε .
t→0,t>0 t t→0,t>0 1
t
for any x ∈ (0, M] because the quantity on the left goes to ∞ when x
becomes closer and closer to 0 (see formula (3.4)). R
∞
The second problem: 2) Is our integral Γ(x) = 0 tx−1 e−t dt uni-
formly convergent on an interval of the type [M, ∞), where M > 0?
The answer is also no. Indeed, for any fixed N > 1,
Z ∞ Z ∞
x−1 −t x−1
t e dt ≥ N e−t dt = N x−1 e−N → ∞,
N N
170 4. INTEGRALS WITH PARAMETERS
Z ∞
= −tx−1 e−t |∞
0 +(x − 1) tx−2 e−t dt = (x − 1)Γ(x − 1)
0
or
(3.5) Γ(x) = (x − 1)Γ(x − 1)
for any x > 1.
3. EULER’S FUNCTIONS GAMMA AND BETA 171
µ ¶µ ¶ √
1 3 1 1 (2n − 1)(2n − 3)...1 π
M2n = n− n− ... Γ( ) = n−1
· .
2 2 2 2 2 2
Let us give some information in order to sketch the graphic of Γ(x)
(see Fig.1)
Theorem
R∞ 61. (derivatives and graphic for Γ(x)) Gamma function
Γ(x) = 0 tx−1 e−t dt, x ∈ (0, ∞) has the following additional proper-
ties:
1) Γ00 (x) > 0, i.e. the graphic of Γ is a convex one (keep water!-see
Fig.1)
2) Γ has a unique local extremum point, namely a minimum point
xmin ∈ (1, 2) (see Fig.1)
3) Oy-axis is a vertical asymptote (see Fig.1)
4) lim Γ(x) = ∞. Moreover, Γ(x) → ∞, whenever x → ∞ as fast
x→∞
as n! → ∞, when n → ∞ (see Fig.1).
R∞
Proof. 1) Formula (3.2) implies Γ00 (x) = 0 tx−1 (ln t)2 e−t dt > 0
for any x ∈ (0, ∞). Thus the function Γ(x) is convex.
2) Since Γ(1) = Γ(2) = 1 (see theorem 60), Rolle’s theorem (see for
instance [Po], Th.36) says that there exists a point xmin ∈ (1, 2) such
that Γ0 (xmin ) = 0. Since Γ00 (x) > 0, xmin is a point of minimum for Γ.
It is the unique extremum point for Γ. Otherwise, if there was another
one z 6= xmin , then Γ0 (z) = 0 (Fermat’s theorem-see for instance [Po],
Th.35). Applying Rolle’s theorem to the function Γ0 (x) on the interval
with ends z and xmin respectively, we would find a point y in this last
interval such that Γ00 (y) = 0, a contradiction, because Γ00 (x) > 0 for
any x ∈ (0, ∞). Thus xmin is the unique extremum point for Γ.
3) Let us compute
Γ(x + 1) 1
lim Γ(x) = lim = = ∞.
x→0,x>0 x→0,x>0 x +0
Here we just used the recurrence formula Γ(x+1) = xΓ(x) (see theorem
60) and the continuity of Γ (see theorem 59).
4) The continuity of Γ, just used above, and the basic formula
Γ(n) = (n − 1)! implies that
lim Γ(x) = lim Γ(n) = lim (n − 1)! = ∞.
x→∞ n→∞ n→∞
¤
3. EULER’S FUNCTIONS GAMMA AND BETA 173
Figure 1
point (x0 , y0 ) ∈ (0, ∞) × (0, ∞) and let us choose p0 , q0 > 0 such that
x0 > p0 and y0 > q0 . Thus, (x0 , y0 ) ∈ [p0 , ∞) × [q0 , ∞) and, since the
beta function B(x, y) is uniformly convergent on [p0 , ∞)×[q0 , ∞), from
theorem 57 b) and d), applied to B0 (x, y) and to B1 (x, y) separately,
we see that B(x, y) is continuous, differentiable and even of class C ∞
on (0, ∞) × (0, ∞) (work slowly everything!). ¤
In the following result we put together the main properties of the
function beta.
3. EULER’S FUNCTIONS GAMMA AND BETA 175
Figure 2
R1
Theorem 63. Let B(x, y) = 0 tx−1 (1 − t)y−1 dt, x > 0, y > 0 be
the beta function. Then
1) B(x, y) = B(y, x) (symmetry).
y
2) B(x, y + 1) = x+y B(x, y) (reduction formula).
x
3) B(x + 1, y) = x+y B(x, y) (reduction formula).
R ∞ ux−1
4) B(x, y) = 0 (1+u) x+y du.
R ∞ uy−1
5) B(x, y) = 0 (1+u)x+y du.
6) B(x, y) = Γ(x)·Γ(y)
Γ(x+y)
(Γ-reduction). This formula reduces the com-
putations with function
Rπ B to computations with function Γ.
2x−1
7) B(x, y) = 2 0 sin
2
u cos2y−1 udu (the trigonometric expres-
sion).
Proof. First of all we want to remark that all the following com-
putations are made in fact with proper integrals, then we take limits
to obtain the improper integrals formulations. But we omit to specify
this and we directly work with the improper form! Let us use this
extremely excessive method only to prove 1).
Z η
B(x, y) = lim tx−1 (1 − t)y−1 dt =
ε→0,η→1 ε
Z 1−η
u=1−t
= − lim (1 − u)x−1 uy−1 du =
ε→0,η→1 1−ε
176 4. INTEGRALS WITH PARAMETERS
Z 1
= uy−1 (1 − u)x−1 du = B(y, x).
0
2)
Z 1 Z 1 µ x ¶0
x−1 y t
B(x, y + 1) = t (1 − t) dt = (1 − t)y dt =
0 0 x
µ x¶ ¯1 Z 1 Z
t ¯ y
y¯ x y−1 y 1 x−1
= (1 − t) ¯ + t (1−t) dt = t (1+t−1)(1−t)y−1 dt =
x 0 x 0 x 0
Z 1 Z 1
y y y y
tx−1 (1 − t)y−1 dt − tx−1 (1 − t)y dt = B(x, y) − B(x, y + 1).
x 0 x 0 x x
Thus h yi y
1+ B(x, y + 1) = B(x, y),
x x
or
y
B(x, y + 1) = B(x, y).
x+y
3) We use now 1), 2) and again 1) to prove 3):
x x
B(x + 1, y) = B(y, x + 1) = B(y, x) = B(x, y).
x+y x+y
R1
4) In B(x, y) = 0 tx−1 (1−t)y−1 dt let us make the change of variable
u
t = 1+u =
u 1
t= =1− .
1+u 1+u
We simply get
Z ∞µ ¶x−1 µ ¶y−1 Z ∞
u 1 1 ux−1
B(x, y) = du = du.
0 1+u 1+u (1 + u)2 0 (1 + u)x+y
5) If in the formula 4) we use the symmetry 1) we get:
Z ∞
uy−1
B(x, y) = B(y, x) = du,
0 (1 + u)y+x
i.e. formula 5).
6) In formula
Z ∞
Γ(p)
(3.8) y p−1 e−ty dy = , t > 0, p > 0.
0 tp
of theorem 60 9) we simply put t + 1 instead of t and p + q instead of
p (here q > 0). We get:
Z ∞
Γ(p + q)
= y p+q−1 e−(t+1)y dy.
(t + 1)p 0
3. EULER’S FUNCTIONS GAMMA AND BETA 177
Let us multiply this last equality by tp−1 and then let us integrate the
obtained result with respect to t from 0 up to ∞:
Z ∞ Z ∞ ∙Z ∞ ¸
tp−1 p−1 p+q−1 −(t+1)y
(3.9) Γ(p + q) dt = t y e dy dt.
0 (t + 1)p+q 0 0
It is not so difficult to verify the conditions of theorem 57 c) and apply
Fubini’s result, i.e. we change the order Rof integration in the equal-
∞ tp−1
ity (3.9). Moreover, we also use formula 0 (t+1) p+q dt = B(p, q) just
obtained in 4):
Z ∞ ∙Z ∞ ¸
p+q−1 −y p−1 −ty
Γ(p + q)B(p, q) = y e t e dt dy.
0 0
We use again formula (3.8) with t instead of y and y instead of t to
compute the brackets inside the right side of the last equality:
Z ∞
Γ(p)
Γ(p + q)B(p, q) = y p+q−1 e−y p dy =
0 y
Z ∞
= Γ(p) y q−1 e−y dy = Γ(p)Γ(q),
0
or
Γ(p)Γ(q)
B(p, q) = .
Γ(p + q)
R1
7) In the definition formula B(x, y) = 0 tx−1 (1 − t)y−1 dt of the
function beta let us put t = sin2 u :
Z π
2
B(x, y) = sin2x−2 u · cos2y−2 u · 2 sin u · cos udu =
0
Z π
2
=2 sin2x−1 u cos2y−1 udu.
0
¤
The function beta is useful in the computation
Rπ of some types of
7 √
integrals. For instance, let us compute I = 0 sin x 4 cos xdx. It is
2
5 3 Γ( 5 )Γ( 3 ) 1
Γ( 1 ) · Γ( 34 )
J = B( , ) = 45 34 = 4 4 .
4 4 Γ( 4 + 4 ) 1
1
Let us use the complementaries √
formula 7) of
√
theorem 60 for x = 4
and
1 3 π π 2 π 2
find Γ( 4 ) · Γ( 4 ) = sin π = 2 . Hence J = 8 .
4
Rπ
Example 63. Another important integrals are I1 (m) = 02 sinm xdx
Rπ
and I2 (m) = 02 cosm xdx. Let us compute I1 (m) (here m is a natural
number).
Z π µ ¶ ¡ ¢ 1
2
m 1 m+1 1 1 Γ m+1 Γ( )
I1 (m) = sin xdx = B , = ¡ m+2 ¢ 2 .
2
0 2 2 2 2 Γ 2
If m = 2k (even), then
£ ¤2
1 Γ(k + 12 )Γ( 12 ) 1 (k − 12 )(k − 32 )... 12 Γ( 12 )
I1 (2k) = = =
2 Γ(k + 1) 2 k!
π · 1 · 3 · 5 · ... · (2k − 1)
= .
2k+1 k!
If m = 2k + 1 (odd), then
1 Γ(k + 1)Γ( 12 ) 1 k!Γ( 12 )
I1 (2k + 1) = = =
2 Γ(k + 1 + 12 ) 2 (k + 12 )(k − 12 )(k − 32 )... 12 Γ( 12 )
2k k!
= .
1 · 3 · 5 · ... · (2k + 1)
We leave as an exercise to the reader to compute I2 (m).
Example 64. ("Traian Lalescu" national contest, 2006, Romania)
Rb
Let us compute Jn = a (x − a)n (b − x)n dx. First of all let us change
the variable x with u = x − a. Thus,
Z b−a Z b−a µ ¶n
n n n n u
Jn = u (b − a − u) du = (b − a) u 1− du.
0 0 b−a
u
Let us change again the variable u with t = b−a . Hence
Z 1
Jn = (b − a)2n+1
tn (1 − t)n dt = (b − a)2n+1 B(n + 1, n + 1) =
0
many times);
R1 2 x2 )
d) 0 ln(1−α
x2 1−x2 dx, |α| < 1;
√
Rπ
e) 0 ln(1+sin a cos x)
dx;
R ∞ e−ax −e−bx
cos x
f) 0 dx, a > 0, b > 0;
R ∞ e−axx−e−bx
g) 0 x
sin(mx)dx, a, b > 0;
R π ¡ 1+a ¢
h) 02 ln 1−a cos xx cos1 x dx, |a| < 1.
cos
Rπ
6. Prove that the famous Bessel’s function Jn (x) = π2 0 cos(nt −
x sin t)dt, n = 0, 1, 2, ... , is a solution of the ordinary differential equa-
tion (Bessel’s equation): x2 y 00 + xy 0 + (x2 − n2 )y = 0.
7. Compute:
U sin x t2 Ux 2 U x2 −t2 U x2 t3
e dt (tan−1 t) dt e dt x e dt
a)lim U 0tan x et2 dt ; b) lim 0 √x2 +1 ; c)lim Uxx2 −t2 ; d)lim Usin x+1 t2
e dt
.
x→π 0 x→∞ x→0 x3
e dt x→0 1
8. RUse Dirichlet’sR integral to compute:
R∞ 2
∞ 3 ∞ 4
a) 0 sinx x dx; b) 0 sinx x dx; c) 0 sinx2 x dx;
Hint: sin 3x = (sin x) (3 − 4 sin2 x); sin4 x = (1 − cos2 x) sin2 x;
9. Use informations on Euler’s gamma and beta functions to com-
pute:
180 4. INTEGRALS WITH PARAMETERS
R∞ 1 Rπ
a) 0 t 4 (1 + t)−2 dt; b) 02 sin3 x cos5 xdx;
Rπ R1 2 1
c) 02 sin3 x cos2 xdx; d) 0 x− 3 (1 − x)− 3 dx;
R∞ 2 R∞ 2
e) 0 x6 e−x dx; f) −∞ xn e−x dx, n a natural number;
R∞ 3 R1 √
g) 0 e−x dx; h) 0 t3 1 − tdt;
R1 R1 1 R∞ x R∞ 1
i) 0 (1 − x2 )10 dx; j) 0 √1−x 3 dx; k) 0 1+x3 dx; l) 0 1+x3
dx;
R ∞ dx R ∞ √3 x
m) 0 (1+x2 )2 ; n) 0 1+x2 dx.
10) Compute: ¯
³R 2 ´00 ¯ Rπ Rπ
dx ¯¯ ; b) 04 ln(1+t·tan x)dx, t ≥ 0; c) 02 ln(cos2 x+
t −tx2
a) t e
√ t=1 Rπ
2 sin2 x)dx (Hint: Compute first of all I(t) = 02 ln(cos2 x+t sin2 x)dx,
√
then put R ∞t = 2);
x2
d) 0 1+x4 dx;
¡ ¢ √
11) Prove that equality: Γ 12 = π implies equality:
¡ ¢ √
Γ n + 12 = 1·3·5·...·(2n−1) n π for any n ∈ N.
¡1¢ 2 ¡ ¢ R∞√ 4
12) If A = Γ 6 and B = Γ 13 , compute 0 6 x(1 + x)− 3 dx as
an expression of A and B.
13) Reduce to computations with function Γ the following integral
Rπ
2
sin10 x cos12 xdx. ¡ ¢
0 √
14) Prove √that the equality: Γ 12 = π implies the equality:
R ∞ −x2
0
e dx = 2π . R∞ R∞
15) Prove that m = −∞ xf (x)dx and σ 2 = −∞ (x − m)2 f (x)dx,
where
1 (x−m)2
f (x) = √ e− 2σ2 ,
σ 2π
where σ > 0 and m is a fixed real number. This last function is called in
Statistics the probability function of the Gauss-Laplace distribution . It
models for instance the distribution of errors in a particular sequences
of random measurements. m is called the mean and σ 2 is called the
variance of this last distribution. Rλ
16) Compute I(λ) = 0 ln(1+λx) 1+x2
dx, |λ| < 1 and then compute
R 0.5 ln(1+0.5x)
0 1+x2
dx.
CHAPTER 5
Line integrals
Figure 1
Figure 2
C will be a finite union of simple curves Ci , i = 1, 2, ..., N, such that
Ci ∩ Cj = ∅ or a point for i 6= j.
Figure 3
Let us come back to the circle (1.2), but instead of the interval
[0, 2π) we consider a new interval [−π, π); this means that we start to
go on the circle from M(−R, 0), in the counterclockwise direction, up
to we come back to M. This is the same with changing the variable t
with a new one v = t − π ∈ K = [−π, π). The new parametrization in
the variable v is:
½
x(v) = −R cos v
, t ∈ [0, 2π).
y(v) = −R sin v
Now we change the variable v ∈ K = [−π, π) with another one u ∈
J = [−∞, ∞) such that tan v2 = u. Thus, the new parametrization of
the same circle is:
½ 2
xe(u) = −R 1−u
1+u 2
(1.4) 2u , u ∈ [−∞, ∞) ⊂ R = R ∪ {±∞}.
ye(u) = −R 1+u2
184 5. LINE INTEGRALS
Figure 4
of the arc [AB]d (see Fig.4). If the norm k∆k of the division ∆ is
small enough then the uniform continuity of f implies that the length
of the arc [M\ i−1 Mi ] are sufficiently small (as we want). Thus we can
approximate the density function f on such an arc with the value of
it at a fixed point Pi (x(ξ i ), y(ξ i ), z(ξ i ) ∈ [M\
i−1 Mi ], where ξ i ∈ [ti−1 , ti ]
for any i = 1, 2, ..., n. The points {ξ i } are called marking points for
the wire (I, − →r , f ) and division ∆. We construct now a new "abstract"
wire which approximate the initial one. Namely, we approximate the
d with the union of the straight line segments [Mi−1 Mi ], i =
arc [AB]
1, 2, ..., n. Then we define a density function on this new curve (a polyg-
onal line) by taking f (x) = f (Pi ) = f (x(ξ i ), y(ξ i ), z(ξ i )) for any x ∈
[Mi−1 Mi ]. The mass of this new obtained wire (the polygonal line,
186 5. LINE INTEGRALS
Definition 17. We say that the general wire (I = [a, b], − →r , f ) has
a mass H (a real number) if for any small positive real number ε > 0,
there exists a small positive real number δ (which depends on ε) such
that |H − S((I, − →r , f ); ∆; {ξ i })| < ε for any division ∆ of [a, b] with
k∆k < δ and for any set of marking points {ξ i }, ξ i ∈ [ti−1 , ti ]. This
is also equivalent to saying that H is a unique point with the follow-
ing property: "There exists a sequence of divisions {∆n }, ∆n ≺ ∆n+1
(the points of ∆n are between the points of ∆n+1 ) such that k∆n k → 0
and the sequence {S((I, − → (n)
r , f ); ∆n ; {ξ i })} is convergent to H for any
(n)
choice of the set of marking points {ξ i } for the division ∆n . This num-
ber H is said to be the line (curvilinear) integral of the firstR type of the
→
−
function f on the curve C = r ([a, b]).It is denoted by C f (x, y, z)ds,
were ds is called the element of length, i.e. the length of a "small"
curvilinear arc [M\ →
−
i−1 Mi ] on C. If aR wire (I = [a, b], r , f ) has a mass
H we also say that the line integral C f (x, y, z)ds exists and it is equal
to H. When f = 1, the mass H coincides with the length l(C). Thus,
in this last case,
R the length exists (C is rectifiable-see definition 8) and
it is equal to C ds.
X
n
f (x(ξ i ), y(ξ i ), z(ξ i ))×
i=1
p
× [x(ti ) − x(ti−1 )]2 + [y(ti ) − y(ti−1 )]2 + [z(ti ) − z(ti−1 )]2 .
X
n
= f (x(ξ i ), y(ξ i ), z(ξ i ))×
i=1
p
x02 (ci )(ti − ti−1 )2 + y 02 (di )(ti − ti−1 )2 + z 02 (ei )(ti − ti−1 )2 ,
where ci , di , ei ∈ [ti−1 , ti ] (we applied Lagrange formula for each func-
tions x(t), y(t) and z(t) on the intervals [ti−1 , ti ]). Since the curve
C is of class C 1 on [a, b], functions x0 (t), y 0 (t) and z 0 (t) are continu-
ous, so that we can use freely the approximations: x02 (ci ) ≈ x0 (ξ i ),
y 02 (di ) ≈ y 0 (ξ i ), z 02 (ei ) ≈ z 0 (ξ i ). Thus, our last Riemann sum becomes:
(1.7) S((I, − →r , f ); ∆; {ξ }) ≈
i
X
n
p
f (x(ξ i ), y(ξ i ), z(ξ i )) x02 (ξ i ) + y 02 (ξ i ) + z 02 (ξ i )(ti − ti−1 ).
i=1
The approximation here is better and better if the norm of the di-
visions ∆ is smaller and smaller. Hence these Riemann sums have
a unique limit point H if and only if the sums on the right in for-
mula (1.7) have such a point H. But the sums on the right are noth-
ing
R b else than Riemann p sums for the usual definite Riemann integral
a
f (x(t), y(t), z(t)) x02 (t) + y 02 (t) + z 02 (t)dt. Thus, finally, this last
integral exists and it is equal to H. Hence H exists and the formula
(1.6) is true. ¤
188 5. LINE INTEGRALS
Figure 5
Figure 6
d = π 4
length([AB]) . ds = dt =
g
[AB] 0 4
Now, the length of [OA] is obviously 1 and the perimeter of the frame
is equal to 1 + π4 + 1 = 2 + π4 .
b) Since
mass[OABO] = mass[OA] + mass[AB] d + mass[OB] =
Z Z Z
2 2 2 2
= (x + y )ds + (x + y )ds + (x2 + y 2 )ds =
[OB] g
[AB] [OA]
Z √1
√ Z π Z 1
2 4
2 2 2 2
= (t + t ) 2dt + (cos t + sin t)dt + (t2 + 02 )dt =
0 0 0
1 π 1 2 π
= + + = + .
3 4 3 3 4
c) First of all let us explain some formulas in connection with the mass
d By
centre. Let G(xG , yG ) be the mass centre of the loaded arc [AB].
definition, if we concentrate the entire mass of the arc in the point G,
both static moments with respect to axes of this last obtained system is
equal to static moments with respect to axes of the initial arc. If one
1. THE MASS OF A WIRE. LINE INTEGRALS OF THE FIRST TYPE. 191
and Z Z
yG · f (x, y)ds = yf (x, y)ds.
g
[AB] g
[AB]
Hence
R R
g
[AB]
xf (x, y)ds [AB] g yf (x, y)ds
(1.14) xG = R , yG = R .
g
[AB]
f (x, y)ds g
[AB]
f (x, y)ds
Now we have to compute three line integrals. But the mass was just
computed in b), mass = π4 . Let us compute the line integrals which
appear at the numerators:
Z Z
xf (x, y)ds = x(x2 + y 2 )ds =
g
[AB] g
[AB]
Z π
4 π 1
= cos t(cos2 t + sin2 t)dt = sin t |04 = √ .
2
Z0 Z
yf (x, y)ds = y(x2 + y 2 )ds =
g
[AB] g
[AB]
Z π
4 π 1
= sin t(cos2 t + sin2 t)dt = − cos t |04 = 1 − √ .
0 2
Hence √ √
√1
2 2 2 4−2 2
xG = = , yG =
π .
π 4
π
d) Recall that the moment of inertia of a point M(x, y) loaded with a
mass m w.r.t. Oy-axis is equal to mx2 (here x is the distance from
M to Oy-axis). By a process of "globalization" i.e. of "integration"
(explain it like in the case of the mass centre!) we get
Z Z √1 ¯1
5 ¯ √2
2 √ √ x 1
IOy = x2 fds = t2 (t2 + t2 ) 2dt = 2 2 ¯¯ = .
[OB] 0 5 0 10
192 5. LINE INTEGRALS
Let us put together all the formulas which give us the coordinates
of mass centre of a space wire (I, −
→
r , f ), where −
→r (t) = (x(t), y(t), z(t)).
Let C be the support curve of this wire. Then,
R R R
xf ds yf ds zfds
(1.15) xG = R C
; yG = R C
; zG = RC .
C
f ds C
fds C
fds
Let now denote Ix , Iy , Iz the moments of inertia of the above wire w.r.t.
Ox-axis, Oy-axis and Oz-axis respectively. Let IO be the moment of
inertia of the same wire w.r.t. the origin O. Then
Z Z
(1.16) Ix = (y + z )fds; Iy = (x2 + z 2 )fds;
2 2
C C
Z Z
Iz = (y 2 + x2 )fds; IO = (x2 + y 2 + z 2 )f ds.
C C
Looking at the definition 17 and at the formula 1.5 we obtain that the
d [A
area σ(ABB 0 A0 ) of the cylindrical surface bounded by [AB], [
0 B 0 ] and
0 0
the segments [AA ], [BB ] (see Fig.7) can be computed by using the
following formula:
Z
0 0
(1.17) σ(ABB A ) = f ds.
g
[AB]
Figure 7
Example 68. Let us find the area σ of the surface which are a part
of the cylindrical surface with directrix curve the parabola y = 2x2 in
xOy-plane, generator lines parallel to Oz-axis, delimited by the xOy-
plane and the plane x + y + z = 10 (see Fig.8). Since z = 10 − x − y,
f (x, y) = 10−x−y in formula 1.17 and since x = t, y = 2t2 , t ∈ [−1, 1]
\ we get:
is a parametrization of the curve [AOD],
Z Z 1 √
σ= (10 − x − y)ds = (10 − t − 2t2 ) 1 + 16t2 dt.
\
[AOD] −1
√
Since the function g(t) = t 1 + 16t2 is an odd function
R 1 on
√ the symmet-
ric interval [−1, 1] w.r.t. the origin O, one has that −1 t 1 + 16t2 dt =
194 5. LINE INTEGRALS
Figure 8
Figure 9
If the norm of the division ∆ is very small, since the curve [AB] [
is smooth, the lengths of the subarcs [M\ i−1 Mi ] are very small. Thus
→
−
we can approximate the work of a continuous field of forces F (x, y, z)
[ along the
= (P (x, y, z), Q(x, y, z), R(x, y, z)) defined on the curve [AB],
196 5. LINE INTEGRALS
→
− X−
→
n
−−−−−→
(2.2) S( F , ∆, {ξ i }) = F (Pi ) · Mi−1 Mi ,
i=1
where Pi is a fixed point on the arc [M\ i−1 Mi ], i.e. Pi (x(ξ i ), y(ξ i ), z(ξ i )),
where {ξ i }, ξ i ∈ [ti−1 , ti ], is a fixed set of marking points for the division
→
−
∆. Such a sum is called a Riemann sum for F , ∆ and {ξ i }. Here the
\ →
− \ →
−
work W−→ ([M
F i−1 Mi ])of F on [Mi−1 Mi ] was approximated with F (Pi ) ·
−−−−−→
Mi−1 Mi .
where ei ∈ [ti−1 , ti ]. Since the norm k∆k is very small, i.e. the lengths
of intervals [ti−1 , ti ] are small enough and since the functions x0 (t), y 0 (t),
z 0 (t) are continuous (the curve is of class C 1 being smooth), we can well
approximate x0 (ci ), y 0 (di ) and z 0 (ei ) with x0 (ξ i ), y 0 (ξ i ) and with z 0 (ξ i )
respectively (ci , di and ei are very close to ξ i and the previous functions
are continuous!). Thus formula (2.5) can be rewritten as
→
− −−−−−→
F (Pi ) · Mi−1 Mi =
= [P (x(ξ i ), y(ξ i ), z(ξ i ))x0 (ξ i ) + Q(x(ξ i ), y(ξ i ), z(ξ i ))y 0 (ξ i )](ti − ti−1 )+
+R(x(ξ i ), y(ξ i ), z(ξ i ))z 0 (ξ i )(ti − ti−1 ).
→
− P − → −−−−−→
Thus, the Riemann sum S( F , ∆, {ξ i }) = ni=1 F (Pi ) · Mi−1 Mi can be
approximated with
X
n
[P (x(ξ i ), y(ξ i ), z(ξ i ))x0 (ξ i ) + Q(x(ξ i ), y(ξ i ), z(ξ i ))y 0 (ξ i )+
i=1
→
− →
− →
−
Example 69. Let F (x, y) = x2 i + xy j be a plane field of forces
(or, equivalently, let ω = x2 dx + xydy be a plane differential form of
order I) and let Γ be the marked with arrows oriented curve in Fig.10
(the arrows indicate the orientation!)
Figure 10
I
d b) Compute −
→
a) Find the work on the arc of the circle [BC]; F ·
[
[AB]
I
d−
→
r ; c) Compute P dx + Qdy;
[
[CA]
→
−
d) Compute the circulation of the field F along the closed oriented
(like in Fig.10) curve [CABC]. Since our last curve is not smooth (it
has "corners"!) we must decompose it into 3 smooth nonoverlaping
[ [AB]
curves: [CA], [ and [BC][ respectively and then to write (see defin-
ition 19):
I I I I
−
→ − → → −
− → → −
− → → −
−
(2.7) F ·dr = F ·dr + F ·dr + F · d→r.
[CABC] [
[CA] [
[AB] [
[BC]
[
Let us find 3 separate parametrizations for the component curves [CA],
[ and [BC].
[AB] [
½
[ x = 2 + (−2)t
[CA] : , t ∈ [0, 1];
y = 0 + (−1)t
200 5. LINE INTEGRALS
½
[: x=0
[AB] , t ∈ [−1, 2];
y=t
½
[: x = 2 cos t π
[BC] , t ∈ [0, ]− ,
y = 2 sin t 2
π
i.e. t goes from to 0. Now we are ready to compute everything. a)
2
[ can be find as follows:
The work on the arc [BC]
I I
→ −
− →
F ·dr = P dx + Qdy =
[
[BC] [
[BC]
Z π
2 £ ¤
=− (2 cos t)2 (2 cos t)0 + (2 cos t)(2 sin t)(2 sin t)0 dt =
0
Z π
2 £ ¤
= −8 − cos2 t sin t + cos2 t sin t dt = 0.
0
b)
I Z
− −
→ 2 £ 2 ¤
F · d→
r = 0 · 0 + 0 · tdt = 0.
−1
[
[AB]
c)
I Z 1 £ ¤ 7
P dx + Qdy = (2 − 2t)2 (−2) + (2 − 2t)(−t)(−1) dt = − .
0 3
[
[CA]
→
−
Proof. Since the second law of dynamics says that F = m− →
v 0 , we
can apply formula (2.3) and successively write:
I Z b Z b
→ −
− → →−
− → →
− →
−
W = F ·dr = F ( r (t)) · v (t)dt = m v 0 (t) · −
→
v (t)dt =
Γ a a
Z b µ→
− ¶0 ¯b
v ·−→
v m → 2¯
=m dt = |−
v | ¯ = T (B) − T (A).
a 2 2 a
¤
Remark
I 20. Maybe somebody asks what happens with a line in-
→ −
−
tegral F · d→r if we change the parametrization of the given curve
Γ
→
−
Γ, i.e. if the parametric path − →r changes with an equivalent one r∗ .
To preserve the initial orientation we must assume that the change of
variables Φ : [a∗ , b∗ ] → [a, b] is an increasing one, i.e. Φ0 (t) > 0. Then,
→∗ ∗
−
r (t ) = −→ ∗
r (Φ(t∗ )) and dt∗ = dtdt dt. Thus,
I Z b∗ " →#
−
→−
− →∗ →∗
− →−
− →∗ d r∗
F (r ) · dr = F ( r ) · ∗ dt∗ =
Γ a∗ dt
Z b∗ ∙ ¸
→−
− → ∗ d−→
r dt
= F ( r (Φ(t ))) · ∗
dt∗ =
a∗ dt dt
Z b I
→−
− →0
− → →
−
= F ( r (t)) r (t)dt = F · d−
→ r.
a Γ
Thus the value of the line integral does not change. It depends only on
the curve itself and not on its particular parametrization.
Figure 11
Indeed, let us find the general form of all the potential functions
→
−
U (x, y, z) of G :
⎧ ∂U
⎨ ∂x = 0
∂U
=0 .
⎩ ∂U∂y
∂z
= −mg
Let us integrate the last equation w.r.t. z and put the constants which
appear only once:
U(x, y, z) = −mgz + K(x, y),
where K(x, y) is a constant w.r.t. z but, in general, it can depend on x
and y because, in general, U may depend on x and y. Let us introduce
204 5. LINE INTEGRALS
K 0 (x) = 0,
U(x, y, z) = −mgz + K.
→
−
where U(x, y, z) is a Ipotential function (a primitive) of F .
→ −
−
b) In particular, F · d→
r does not depend on the curve Γ which
Γ I
→ →
−
connect the points A and B, i.e. the integral F · d− r does not depend
on path in D.
−
→
c) Another consequence is that two primitives U and V of F on D
differ one to each other by a constant only.
3. INDEPENDENCE ON PATH. CONSERVATIVE FIELDS 205
→
− →
−
I d) For any "closed" curve δ ( r (b) = r (a)) of D one has that
→ −
−
F · d→
r = 0. Conversely, if for any smooth closed curve δ of D we
δ I I
→ −
− → →
−
have that F · d r = 0, then F · d−
→ r does not depend on path in D.
δ
Figure 12
Let us compute the circulation (or the work) of the plane field
−
→ →
− →
−
F (x, y) = 2xy i + x2 j along the curve γ. Because the curve is a
union of 4 smooth distinct curves, we ought to find a parametrization
for each of them, compute 4 line integrals and then add the four ob-
tained numbers (see example 69). All of this may be very boring and
3. INDEPENDENCE ON PATH. CONSERVATIVE FIELDS 207
tiresome. The best method is to check first of all if the field is conserv-
ative or not. If we were lucky that the field to be conservative, then the
line integrals does not depend on path and we can choose a very easy
path which connect the starting and the end points of the arc. In our
case the arc is closed, so that the integral is zero. Let us check now if
our field is conservative or not. For this, let us try to find U(x, y) such
that
∂U ∂U
= 2xy, and = x2 .
∂x ∂y
We integrate the first equality w.r.t. x and find: U = x2 y+K(y). Let us
put this expression of U in the second equation and find: x2 + K 0 (y) =
x2 . Thus, K 0 (y) = 0 and K(y) is a constant K. Hence U(x, y) =
x2 y + K, where K is an arbitrary constant. Therefore our field is
conservative and, as a consequence, the given line integral is zero (see
theorem 67).
But, how to test if a field is conservative or not without direct
checking if a potential (primitive) function exists, like we did above?
We begin with the following remark.
→
−
Theorem 68. Let F : D → R3 ,
→
−
F (x, y, z) = (P (x, y, z), Q(x, y, z), R(x, y, z)) be a conservative field
→ −
− → →
−
of class C 1 on a space domain D. Then curl F = 0 , i.e. the field F
→
− →
−
is an irrotational field. If F : D → R2 , F (x, y) = (P (x, y, ), Q(x, y, )),
is a plane conservative field of class C 1 on a plane domain D, then
∂Q
∂x
= ∂P
∂y
on D.
Proof. Let U (x, y, z) be a primitive function for the conservative
field
→
−
F (x, y, z) = (P (x, y, z), Q(x, y, z), R(x, y, z)).
Then ∂x = P, ∂U
∂U
∂y
= Q and ∂U ∂z
= R. We know that
µ ¶ µ ¶ µ ¶
→
− →
− ∂R ∂Q − → ∂R ∂P − → ∂Q ∂P − →
curl F = ∇× F = − i− − j+ − k.
∂y ∂z ∂x ∂z ∂x ∂y
If here instead of P, Q, R we put ∂U , ∂U and ∂U
∂x ∂y ∂z
respectively, we get
∂R ∂Q ∂2U ∂2U
∂y
− ∂z = ∂y∂z − ∂z∂y = 0 because U is a function of class C 2 (since
P, Q, R are functions of class C 1 ) and we can apply Schwarz’ theorem
(see theorem 71, [Po]). Similarly we can easily show that ∂R ∂x
− ∂P
∂z
=0
∂Q ∂P →
− →
−
and ∂x − ∂y = 0, i.e. curl F = 0 . In the plane field situation, we put
∂2U ∂2U
instead Q, ∂U
∂y
and instead of P, ∂U
∂x
, so that ∂Q
∂x
= ∂x∂y and ∂P∂y
= ∂y∂x .
2 ∂2U ∂2U ∂Q ∂P
Since u is of class C , ∂x∂y = ∂y∂x , i.e. ∂x = ∂y . ¤
208 5. LINE INTEGRALS
x ∂Q y2 −x2 ∂P
and Q = x2 +y 2
. It is easy to see that ∂x
= (x2 +y 2 )2
= ∂y
but
I Z 2π
y x
− 2 2
dx + 2 dy = [sin2 t + cos2 t]dt = 2π 6= 0.
x +y x + y2 0
x2 +y2 =1
→
−
Hence F cannot be conservative on D (see theorem 67 d)).
I
This last remark is very useful in practice. Let us compute xdx +
γ
ydy + zdz, where γ is the oriented path from Fig.13.
→
− →
− →
−
Since F = (x, y, z) is irrotational, i.e. curl F = 0 , the integral
does not depend on path and its value is U (1, 5, 3) − U(0, 0, 0), where
→
−
U is a primitive of F . We know that such a primitive exists from the
last remark (D is the entire R2 so it is simple connected). Thus we can
directly compute it by integrating the system
⎧ ∂U
⎨ ∂x = x
∂U
∂y
=y .
⎩ ∂U
∂z
=z
2 2 2
It is easy to see that x +y2 +z + K, where K is a constant, is the
general solution of this system (use the same idea like in the case of the
3. INDEPENDENCE ON PATH. CONSERVATIVE FIELDS 209
Figure 13
x2 +y2 +z 2
gravitational field). Take U (x, y, z) = 2
and compute U(1, 5, 3)−
U (0, 0, 0) = 35
2
. I
− −
→
F · d→
When a line integral of the second type r does not depend
I
→ −
−
on path on a domain D, then instead of writing F · d→r , where
g
[AB]
d is a path in D which connects two points A and B of D, we
[AB]
I B
→ −
−
simply write F · d→
r . To compute such an integral we can use a
A
→
−
primitive of F or we can use any arc of a piecewise smooth curve γ
included
I B in D which connect A and
I BB in this order.
I Namely, we write
→ −
− →
− → −
−
F · d→r = U(B) − U (A) or F · d−
→r = F · d→
r . The usual
A A γ
choice for γ is a segment of a line or a finite union of segments of lines,
eventually parallel with the coordinate axes. For instance,
I
(3,0,−2) I
(3,−1,2)
I
(3,0,2) I
(3,0−2)
Remark 22. The idea coming from this last example can be gen-
eralized in order to construct a primitive for a conservative continuous
→
− →
−
field F = (P, Q, R) on a domain D. In order to prove that a field F
is conservative on a simple connected domain, the best is to use the
→ −
− →
criterion just described in remark 21 namely, to verify if curl F = 0
(in the space case), or to verify that ∂Q
∂x
= ∂P∂y
(in the plane case). Let
us describe an effective construction for such a primitive U(x, y) in the
plane case (for the space case the description is more tiresome). First
of all let us choose a fixed point M0 (x0 , y0 ) in D. The primitive U we
are going to construct will depend on the choice of this fixed point. We
just know that all the primitives of a given field differ one to each other
by a constant. Thus, such a primitive is completely determined by its
value at a point. In our case we shall determine the primitive U (x, y)
with the property: U (x0 , y0 ) = 0. Take now another point M(c, d) of
D and try to define the value of U at this "moving" point M. Let
[M0 M1 M2 ....M2n−1 M2n ], M2n = M, be a polygonal line in D, which
connect M0 with M, such that the couples of segments
([M0 M1 ], [M1 M2 ]), ([M2 M3 ], [M3 M4 ]), ..., ([M2n−2 M2n−1 ], [M2n−1 M2n ])
has the following property: in each couple ([M2j−2 M2j−1 ], [M2j−1 M2j ]),
j = 1, 2, ..., n, the first segment [M2j−2 M2j−1 ] is parallel to Ox-axis and
the second [M2j−1 M2j ] is parallel to Oy-axis. We start with the value of
U at M0 and shall find the value of it at M2 (we operate with the first
couple). Then we use the value of U at M2 and, by the same procedure,
we find the value of U at M4 and so on up to we succeed to find the
value of U at M2n = M. Hence we reduced everything in describing the
procedure only for the first couple ([M0 M1 ], [M1 M2 ]). More exactly, it is
sufficient to construct U (x, y) in a small open disc B(M0 , r) ⊂ D, r >
0, with centre at M0 and radius r. Let P (s, u) ∈ B(M0 , r) and let the
couple of segments ([M0 P1 ], [P1 P ]) such that P1 (s, y0 ), i.e. the segment
[M0 P1 ] is parallel to Ox-axis and [P1 P ] is parallel to Oy-axis. It is
clear that the polygonal line [M0 P1 P ] is contained in the disc B(M0 , r),
i.e. it is contained in D too. Let us define now:
IP I I
U(s, u) = P dx + Qdy = P dx + Qdy + P dx + Qdy.
M0 [M0 P1 ] [P1 P ]
4. COMPUTING PLANE AREAS BY LINE INTEGRALS 211
or
Z s Z u
(3.3) U (s, u) = P (x, y0 )dx + Q(s, y)dy.
x0 y0
We see that here we have integrals with parameters. Let us use the
→
−
conservativeness hypothesis of F = (P, Q) which implies ∂Q ∂x
= ∂P
∂y
and the Leibniz formula (1.9), Ch.4, to compute (the existence is also
assured!) ∂U∂s
(s, u) and ∂U
∂u
(s, u) :
Z u Z u
∂U ∂Q ∂P
(s, u) = P (s, y0 ) + (s, y)dy = P (s, y0 ) + (s, y)dy =
∂s y0 ∂s y0 ∂y
Figure 14
Z b Z b
= g(x)dx − f (x)dx = area(D).
a a
¤
Remark 23. If the domain D is simple w.r.t. Oy-axis, i.e. if [c, d]
is the projection of D on Oy-axis and if y0 ∈ (c, d), then the straight
4. COMPUTING PLANE AREAS BY LINE INTEGRALS 213
Figure 15
integral is zero (dx = 0). In general, these dividing curves can not be
always chosen to be segments parallel with the Oy-axis. Hence, these
I I
→ −
− → →
−
last two equalities come from the general relation: F ·dr + F ·
γ γ−
d→
−r = 0. Let us use these observations to prove that formula (4.1) is
true even in the general case when the domain is not simple w.r.t. to
Ox-axis but it can be divided into a finite set of nonoverlaping simple
domains w.r.t. Ox-axis. Look at Fig.15 and follow the next reasoning:
I I I
area(D) = area(D1 )+area(D2 )+area(D3 ) = − ydx− ydx− ydx =
∂D1 ∂D2 ∂D3
I I I
= −[ ydx + ydx + ydx+
[A\
4 A3 A2 ] \
[A2 A6 ] \
[A6 A4 ]
I I I I
+ ydx + ydx + ydx + ydx] =
[A\
6 A5 A4 ] \
[A4 A6 ] [A\
2 A1 A6 ] \
[A6 A2 ]
I I I I
− ydx − ydx − ydx = − ydx.
[A\
4 A3 A2 ] [A\
2 A1 A6 ] [A\
6 A5 A4 ]
∂D
In this way we can extend all the formulas (4.1), (4.3) and (4.4) to the
case when the domain D has its boundary ∂D a closed direct oriented
4. COMPUTING PLANE AREAS BY LINE INTEGRALS 215
This last expression of the area does not depend on the origin O or
of the coordinate system. Thus, a very nice consequence of the vector
formula 4.5 is the following.
216 5. LINE INTEGRALS
Figure 16
Figure 17
→
−
f3 (t) k , t ∈ [0, 1], where
⎧
⎨ f1 (t) = a + (x − a)t
f2 (t) = b + (y − b)t , t ∈ [0, 1].
⎩ f (t) = c + (z − c)t
3
The computations are easier in this case (do them!). Finally we get the
same function U defined on D (if A is fixed). For instance, let us use
this last idea to find a primitive of the form ω = yzdx + zxdy + xydz.
→ −
− →
Since curl F = 0 , there exists a primitive (it is clear that D = R3 is
a starred domain) U (x, y, z). Let us fix a centre A(0, 1, 2) (see Fig.18)
Figure 18
and let us compute the value of U at B(3, 1, −3) :
A1I
(3,1,2) I
B(3,1,−3)
Figure 19
7. Let A(a, 0), B(0, b), n = [AB], be the segment AB and let m =
d 2 2
[AB], be the arc of the ellipse xa2 + yb2 = 1 which connects A and B.
The metallic frame [AmBnA] is loaded with a density function f (x, y)
at the point M(x, y), which is three times the value of the projection
of M on the Ox-axis. Find its mass centre.
8. Let A(1, 1), B(1, −1) and [AOB][ be the arc of the parabola x =
[ ∪ [AB] with the density function f (x, y) = x + 1
y 2 . The frame [AOB]
rotates around Ox-axis. Find its moment of inertia.
9. Let A(1,
R 1) and B(2, 0) be two points in the xOy-plane.
Find I = [OA]∪[AB] (x + y 3 )ds.
3
R
R where
10. Find C xyds, C is the square |x| + |y| = a > 0.
√ ds
11. Find I = C 2 2
, where C is the segment which connect
x +y +4
the points O(0, 0) and A(1, 2).
R 2 2
12. Find C xyds, where C is the⎧curve xa2 + yb2 = 1, x, y ≥ 0.
R ⎨ x = 2 cos t
ds
13. Find C x2 +y2 +z2 , where C : y = 2 sin t , t ∈ [0, 4π].
⎩ z = bt
14. Find the length of the arc of a conic helix: x = 2et cos t,
y = 2et sin t, z = 2et , between O(0, 0, 0) and A(2, 0, 2).
15. Find the length and the mass centre of the arc of a cycloid:
½
x = a(t − sin t)
, t ∈ [0, π].
y = a(1 − cos t)
x2 y2
16. Find the mass of the ellipse-wire a2
+ b2
= 1 if its density
function is f (x, y) = |y| . ⎧
R ⎨ x = t2
17. Find C (x + y)ds, where C is the arc: y = 3t
√ , t ∈ [0, 1].
⎩ 2
z = t3
18. Use Riemann sums to find the side area of the parabolic cylinder
y = 38 x2 , bounded by the planes z = 0, x = 0, z = x, y = 6.
I
19. Compute xdy, where C is the direct oriented (trigonometric
C
sense) curve which bounds the triangle realized by the coordinates axes
and the line x2 + y3 = 1.
I
20. Compute xydx + x2 dy, where C is the arc of the parabola
C
y = 4x2 between O(0, 0) and A(1, 4), clockwise oriented.
6. PROBLEMS AND EXERCISES 223
→
− →
− →
−
21. Find the work of the field F (x, y) = xy i + x2 j on the fol-
lowing oriented curves:
a) the rectangle frame [OABCO], where A(0, 1), B(2, 1) and C(2, 0).
The orientation is O → A → B → C →).
b) the curve [OBAO], where [OB] and [AO] are segments of lines,
[BA] is an arc of a circle of radius 2, B(2, 0) and the angle ](AOB) is
equal to 60◦ . The orientation is O → B → A → O.
c) the triangle frame [ABCA], where A(1, 2), B(3, 4) and C(5, 1).
The orientation is A → B → C → A.
d) the curve [OmAnO], where A(1, 1), [OmA] is the arc of the
parabola y = x2 and [AnO] is the arc of the parabola x = y 2 . The
orientation isIO → A → O.
→ →
− →
− →
− →
−
22. Find F · d− r , where −
→r (t) = t i + t2 j + t3 k , t ∈ [0, 1] and
C
−
→
F (x, y, z) = (y, z, 2x). The orientation of C is the direct orientation
induced by the moving of t from 0 to 1.
→
− →
− →
− →
−
23. Find the work of the field F (x, y, z) = xy i + xz j + yz k on
the polygonal frame [ABCO], where A(3, 0, 0), B(3, 3, 0) and C(0, 3, 3).
The orientation is A → B → C → O. Is this field a conservative field?
If it is so, compute again easier the work of the field!
I
24. Find α ∈ R such that the integral I(α) = (αxy−1)dx+3x2 dy
I
(2,3)
does not depend on path. Then compute (αxy − 1)dx + 3x2 dy for
(1,1)
this last found value of α.
25. Prove that the following differential forms are exact on some
simple connected domains (describe them!) and then find their primi-
tives.
a) dx+dy+dz
x+y+z
; b) xdx+ydy+zdz xdy+ydx
√ 2 2 2 ; c) √ .
x +y +z x2 +y 2
I I I
1
26. Use the formulas: area(D) = xdy = − ydx = 2 xdy −ydx
∂D ∂D ∂D
in order to find the areas of the figures bounded by the following closed
curves: ½ ½
2 y2 x = a cos3 t x = t − sin t
a)x + 4 = 1; b) , t ∈ [0, 2π]. c) ,t ∈
y = a sin3 t y = 1 − cos t
[0, 2π], and the Ox-axis.
27. Find the length and the area bounded by
224 5. LINE INTEGRALS
½
x = R(2 cos t − cos 2t)
Γ: , t ∈ [0, 2π], R > 0.
y = R(2⎧cos t − sin 2t)
⎨ x = t cos t
28. Let Γ : y = t sin t , t ∈ [0, π] be a wire with the density
⎩ z=t
function f (x, y, z) = x2 + y 2 . Find its static moment relative to xOy-
plane.
29. Let Γ = {M(x, y) : x2 + y 2 = a2 and x + y + z = 0} with its
direct orientation induced by the direct orientation of the plane xOy.
→
−
Compute the work of F = (1, x, z) on Γ.
30. Let P1 be the parabola x =√y 2 and let√P2 be the unique parabola
which contains the points B(3, − 3), A(3, 3) and C(2, 0). Find the
area bounded by these parabolas.
31. Compute:
R (2,−1) R (−2,1,5) √ √
a) (−1,1) x(1 + x)dx − y(1 + y)dy; b) (0,0,0) x2 dx + ydy − z zdz;
32. Find V (x, y) if dV = (4x + 2y)dx + (2x − 6y)dy. Use this to
solve the differential equation y 0 = 4x−6y
6y−2x
.
CHAPTER 6
Double integrals
Figure 1
and {dj } are sets of marking points for ∆x and ∆y respectively. Thus
we can well approximate:
Xn Xm
def
(1.1) mass(D) ≈ f (ci , dj )area(Dij ) = Sf (∆x , ∆y , {(ci , dj )})
i=1 j=1
These last formulas says that either we integrate f (x, y) w.r.t. y from
c to d and obtain a new function of x which is finally integrated from
a to b (the first equality), or we firstly integrate f (x, y) w.r.t. x from a
to b and obtain a new function of y which is finally integrated from c
to d (the second equality).
Proof. We preserve the above notation in the next considera-
ZZ
R b ³R d ´
tions. Let J = f (x, y)dxdy, I1 = a c f (x, y)dy dx and I2 =
R d ³R b ´D
c a
f (x, y)dx dy. The number J exists because of theorem 73. The
Rd
integral I(x) = c f (x, y)dy is an integral with a parameter x. Thus
Rb
I1 = a I(x)dx and we can well approximate it with Riemann sums of
228 6. DOUBLE INTEGRALS
Z 2à ¯1 ! Z 2µ ¶ µ 2 ¶¯2
x3 x2 2 ¯¯ y y2 y y 3 ¯¯ 5
y + y ¯ dy = + dy = + ¯ = .
1 3 2 0 1 3 2 6 6 1 3
X
n X
m
ci f (ci , dj )(yj − yj−1 )(xi − xi−1 ) = Sxf (∆x , ∆y , {(ci , dj )}),
i=1 j=1
a Riemann sum of function xf (x, y). Hence the theoretical static mo-
ment My of the lamina (D, f ) relative to Oy-axis can be computed by
the formula:
ZZ
(1.5) My = xf (x, y)dxdy.
D
In the same manner we can deduce a formula for the static moment
Mx of the lamina (D, f ) w.r.t. Ox-axis:
ZZ
(1.6) Mx = yf (x, y)dxdy.
D
The coordinates (xG , yG ) of the mass centre G of the lamina (D, f ) are
defined such that if we concentrate the mass of D at the point G, then
both moments of this last system must coincide with the correspondent
global moments, i.e.
ZZ
My = xf (x, y)dxdy = xG · mass(D),
D
and
ZZ
Mx = yf (x, y)dxdy = yG · mass(D).
D
230 6. DOUBLE INTEGRALS
For instance, let us compute the mass centre of the square D = [−1, 1]×
[0, 2] with the density function f (x, y) = y. First of all let us remark
that D is symmetric w.r.t. Oy-axis. Moreover, the distribution of the
mass is symmetric w.r.t. the same axis: f (x, y) = f (−x, y) = y. Thus,
G is on Oy-axis, i.e. xG = 0. In order to compute yG we need to
compute two integrals:
ZZ ZZ Z 1 µZ 2 ¶ Z 1
f (x, y)dxdy = ydxdy = ydy dx = 2 dx = 4
−1 0 −1
D D
and
ZZ ZZ Z 1 µZ 2 ¶ Z 1
2 2 8 16
yf (x, y)dxdy = y dxdy = y dy dx = dx = .
−1 0 3 −1 3
D D
16
Hence yG = 43 = 43 . We leave as an exercise to the reader to prove
that yG = 43 is the mass centre of the wire x = 0, y = t ∈ [0, 2],
with the same density function f (x, y) = y. Use your "static" feeling
to anticipate this result and to explain this feeling.
Example 74. Let us deduce the moments of inertia IOx of a simple
rectangle D with density function f (x, y), when D moves around Ox-
axis. For a material point P (x, y) with a mass m, this moment of
inertia is equal to y 2 m, i.e. the mass m multiplied by the square of the
distance from P to the rotation axis Ox. Thus, we can well approximate
the moment of inertia of the rectangle Dij w.r.t. Ox-axis with the
quantity d2j f (ci , dj )area(Dij ). Thus the global moment of inertia of D
relative to Ox-axis can be well approximated by
Xn X m
d2j f (ci , dj )(yj − yj−1 )(xi − xi−1 ) = Sy2 f (∆x , ∆y , {(ci , dj )}).
i=1 j=1
Figure 2
Thus we have:
ZZ Z µZ ¶ ¯1
2
1
2
1
y 3 ¯¯ 4
IOx = y dxdy = y dx dy = 2 ¯ = .
−1 −1 3 −1 3
D
Figure 3
X
n X
m
f (ci , dj )(yj − yj−1 )(xi − xi−1 ),
i=1 j=1
ZZ
vol = f (x, y)dxdy.
Ω
2. DOUBLE INTEGRALS ON AN ARBITRARY BOUNDED DOMAIN 233
When f (x, y) has also negative values, we can use the more general
formula:
ZZ
(1.10) vol = |f (x, y)| dxdy.
Ω
For instance, in Fig.4 we have a rectangular cylindrical body
[ABCOA0 B 0 C 0 O0 ]. In this case the parallelepiped solid with the basis
the rectangle [ABCO] ¡ is xcuted ¢by the plane xa + yb + zc = 1 in the upper
part. Hence z = c 1 − a − yb , a > 1, b > 1 and the volume of our
cylindrical body [ABCOA0 B 0 C 0 O0 ] can be computed as follows:
Z 1 µZ 1 h ¶ Z 1h
c c i c c i¯¯1
vol = c − x − y dy dx = cy − xy − y 2 ¯ dx =
0 0 a b 0 a 2b 0
Z 1³ ´ ³ ´¯
c c c c ¯1 c c
c− x− dx = cx − x2 − x ¯ = c − − .
0 a 2b 2a 2b 0 2a 2b
Figure 4
For instance, the diameter of a closed disc B[A, r] is 2r, i.e. its usual
diameter. The diameter of a rectangle Ω is equal to the length of one
of its diagonal. The diameter of a parallelogram is equal to the length
of its greatest diagonal. And so on! A division ∆ of the domain D
(like above!) is a finite set of domains Di , i = 1, 2, ..., n, of the same
type like D, such that ∪Di = D and Di ∩ Dj = ∅ or a figure of
zero area (in fact points or piecewise smooth curves!) for i 6= j (see
2. DOUBLE INTEGRALS ON AN ARBITRARY BOUNDED DOMAIN 235
Figure 5
This last result says that the class of integrable functions is included
in the class of bounded functions. This is why any function considered
by us in the following is assumed to be bounded.
Let D be like above, let ∆ = {Di }, i = 1, 2, ..., n be a division of D
and let f be a bounded function defined on D with values in R. Let
mi = inf{f (x, y) : (x, y) ∈ Di } and let Mi = sup{f (x, y) : (x, y) ∈ Di }
for any i = 1, 2, ..., n. Then
X
n
s∆ (f ) = mi · area(Di )
i=1
is called the inferior Darboux sum of f relative to ∆ and
X
n
S∆ (f ) = Mi · area(Di )
i=1
is said to be the superior Darboux sum of f relative to ∆. It is clear
that
m · area(D) ≤ s∆ (f ) ≤ Sf (∆, {Pi (ci , di )}) ≤ S∆ (f ) ≤ M · area(D),
for any set of marking points {Pi (ci , di )} of the division ∆, where m =
inf{f (x, y) : (x, y) ∈ D} and M = sup{f (x, y) : (x, y) ∈ D}. Since the
set {s∆ (f ) : ∆ goes on the set of all divisions of D} is upper bounded by
M · area(D), it has a least upper bound I∗ (f ). Since the set {S∆ (f ) : ∆
goes on the set of all divisions of D} is lower bounded by m · area(D),
it has a greatest lower bound I ∗ (f ). In general, I∗ (f ) ≤ I ∗ (f ).
Theorem 76. (Darboux criterion) Let D be like above and let f :
D → R be a bounded function. Then f is integrable on D if and
only if I∗ (f ) = I ∗ (f ), i.e. if and only if for any ε > 0 there exists a
δ ε > 0 such that if k∆k < δ ε then S∆ (f ) − s∆ (f ) < ε. In this last case
ZZ
∗
I∗ (f ) = I (f ) = f (x, y)dxdy.
D
X
n
¯ ¯ ε X n
≤ ¯f (x0(i) ) − f (x(i) )¯ · area(Di ) < area(Di ) = ε,
i=1
area(D) i=1
° °
because k∆k < δ implies diam(Di ) < δ and so, °x0(i) − x(i) ° < δ.
Thus we can apply the uniform continuity of f and finally we get that
S∆ (f ) − s∆ (f ) < ε, i.e. that f is integrable. ¤
Remark 27. Up to now we used closed bounded domains Di with
piecewise smooth boundaries, as subdomains of a division ∆ of a fixed
domain D of the same type. Since any such subdomain Di can be
well approximated with elementary figures (finite union of nonoverlap-
ing simple rectangles) or with finite unions of nonoverlaping triangles,
discs, parallelograms, etc., we can substitute these Di in definition 24 or
in theorem 76 with triangles, discs, parallelograms, etc. It is clear that
in general it is not possible to cover a domain D with nonoverlaping
discs, for instance. But, for any small ε > 0, there exists such a union
of discs Uε ⊂ D, such that area(DÂUε ) < ε. Hence we can work with
the following alternative equivalent definition for an integrable function
f. We say that f : D → R is (Riemann) integrable on D if there ex-
ists a real number I such that for any small ε > 0 and η > 0 there
exists a union Uη = {Bi } of Pnnonoverlaping discs, such that Uη ⊂ D,
area(DÂUη ) < η and |I − i=1 f ((ci , di ))area(Bi )| < ε for any set of
marking points {Pi (ci , di )}, Pi (ci , di ) ∈ Bi , i = 1, 2, ..., n. Similarly we
can work with triangles, parallelograms, general rectangles, etc., instead
of discs.
238 6. DOUBLE INTEGRALS
The following result put together some basic properties of the dou-
ble integral.
ZZ
Theorem 78. a) the mapping f à f (x, y)dxdy is a linear
D
mapping defined on the vector space Int(D) of all integrable functions
defined on D. This means that
ZZ ZZ ZZ
[αf (x, y) + βg(x, y)]dxdy = α f (x, y)dxdy + β g(x, y)dxdy.
D D D
b) the mass of a lamina (D, f ) can be computed as follows:
ZZ
mass(D) = f (x, y)dxdy.
D
ZZ
In particular the area of D is equal to dxdy.
D
ZZ ZZ
c) if f ≤ g on D, then f (x, y)dxdy ≤ g(x, y)dxdy. In par-
ZDZ D
Figure 6
240 6. DOUBLE INTEGRALS
With these notation and definitions one has the following basic
result.
Theorem 79. (iterative general formula) Let D be a simple domain
w.r.t. Ox-axis like above and let f : D → R be a piecewise continuous
function (bounded and continuous on D except a subset of zero area)
defined on D. Then
ZZ Z b ÃZ β(x) !
(2.2) f (x, y)dxdy = f (x, y)dy dx.
a α(x)
D
R β(x)
Here I(x) = α(x) f (x, y)dy is an integral with a parameter x in its
general form (see formula (1.2)). This formula says that in order
to compute a double integral on this particular type of domain D, we
fix an x on [a, b], the projection of D on Ox-axis, cut the domain D
with the vertical line which passes through x and obtain the segment
[U V ], where U(x, α(x)) and V (x, β(x)), compute the simple integral
R β(x)
I(x) = α(x) f (x, y)dy with a parameter x, along this segment and fi-
Rb
nally compute the next simple integral a I(x)dx. In fact, we just com-
puted the mass I(x) of the segment [UV ], which depends on x, then we
Rb
"made the sum" a I(x)dx of all these masses I(x).
Proof. We intend to use the similar iterative formula for a rec-
tangle (see formula (1.2)) for an extension-function fe(x, y) : Ω =
[LMNP ] → R of f : D → R to Ω. We simply extend f from D to
the rectangle Ω = [a, b] × [c, d] by putting zero in all the points of
ΩÂD :
½
e f (x, y), if (x, y) ∈ D
f (x, y) = .
0, if (x, y) ∈ ΩÂD
Since this function is continuous on Ω except maybe the points of ∂D,
which has zero area, we can apply theorem 77 and find that fe is inte-
grable on Ω and
ZZ ZZ Z Z ZZ
f (x, y)dxdy = fe(x, y)dxdy − 0 · dxdy = fe(x, y)dxdy.
D Ω ΩÂD Ω
But now we can apply the iterative formula (1.2) for the rectangle
Ω = [a, b] × [c, d] and find:
ZZ Z b µZ d ¶
e
f (x, y)dxdy = e
f (x, y)dy dx =
a c
Ω
2. DOUBLE INTEGRALS ON AN ARBITRARY BOUNDED DOMAIN 241
Z b ÃZ α(x) Z β(x) Z d
!
= fe(x, y)dy + fe(x, y)dy + fe(x, y)dy dx =
a c α(x) β(x)
Z b ÃZ β(x)
!
= f (x, y)dy dx.
a α(x)
¤
For instance, let us compute the mass of the plate D with the
density function f (x, y) = 3xy 2 , where D is the domain bounded by
the line y = x and the parabola y = x2 . Since x ∈ [0, 1] and since
x ≥ x2 , when x ∈ [0, 1], we get:
ZZ Z 1 µZ x ¶ Z 1
¯x
mass(D) = 2
3xy dxdy = x 2
3y dy dx = x · y 3 ¯x2 dx =
0 x2 0
D
Z ¯1
1
3 6 x5 x8 ¯¯ 1 1 3
x · (x − x )dx = − ¯ = − = .
0 5 8 0 5 8 40
Remark 28. If the domain D is not simple w.r.t. Ox-axis but it
is simple w.r.t. Oy-axis (see Fig.7) we obtain the analogous formula:
ZZ Z d ÃZ γ(y) !
(2.3) f (x, y)dxdy = f (x, y)dx dy,
c δ(y)
D
Figure 7
Z 3 5
¯
5 √ (x − 1) 2 ¯¯ 4 3 32
=2 x − 1dx = 2 3 ¯ = · 42 = .
1 2
¯ 3 3
1
Let us compute now the static moment w.r.t. Oy-axis, by using formula
(2.3) (decomposing D into two nonoverlaping subdomains, compute
also this moment by using formula (2.2)):
ZZ Z ÃZ !
2 y 2 +1
MOy (D) = xdxdy = xdx dy =
−2 1
D
Z 2 Z 2
1 2 2 24 · 11
= [(y + 1) − 1]dy = [y 4 + 2y 2 ]dy = .
2 −2 0 15
Thus,
ZZ
xdxdy
24 ·11
D 15 33
xG = = 32 = .
mass(D) 3
30
Prove that this xG is the same like the x-coordinates of the mass centre
of the domain D0 bounded by y = 2, x = y 2 + 1 and by x = 1.
Let us now compute the moment of inertia IOx (D) of the plate D
from Fig.8 with the density function f (x, y) = 2x. The formula is the
same formula for rectangles (see formula (1.8)) extended to a general
2. DOUBLE INTEGRALS ON AN ARBITRARY BOUNDED DOMAIN 243
Z " √ #¯2
2 2 √ 3 2 32 2 5 x7 ¯¯ 768
= (16 2x 2 − x6 )dx = x2 − ¯ = .
3 0 3 5 7 ¯ 35
0
Figure 8
Z 1Ã ¯ !
3 ¯1−x Z 1µ ¶
2 y ¯ 2 (1 − x)3
x y+ ¯ dx = x (1 − x) + dx =
0 3 0 0 3
∙ 3 ¸¯1
x x4 (1 − x)4 ¯¯ 1 1 1 1
− − ¯ = − + = .
3 4 12 0 3 4 12 6
Exercise 8. Let D be a simple domain w.r.t. Ox-axis and let [a, b]
be its projection on Ox-axis. For any x ∈ [a, b] let Mx (x, α(x)) and
Nx (x, β(x)), α(x) ≤ β(x) be the points at which the vertical line X = x
cuts ∂D, the boundary of D. Let G(x, θ(x)) be the mass centre (the
density function is 1) of the segment [Mx Nx ]. Let (γ, f ) be the wire
x = x, y = θ(x), and f (x, θ(x)) = length[Mx Nx ], where x ∈ [a, b] and
let H be its mass centre. Prove that H is the mass centre of D.
Exercise 9. State and comment a Lebesgue type criterion for dou-
ble integrals.
Exercise 10. Let f, g : D → R be two piecewise continuous func-
tions defined on a domain D like above. Let A = {(x, y) ∈ D :
f (x, y) 6= g(x, y)}. We assume that A has zero area. Prove that
ZZ ZZ
f (x, y)dxdy = g(x, y)dxdy.
D D
Figure 9
½
x=x
(3.3) arc[BDA] : , x ∈ [a, b]− ,
y = ψ(x)
½
x = α(y)
(3.4) arc[DAC] : , y ∈ [c, d]− ,
y=y
½
x = β(y)
(3.5) arc[CBD] : , y ∈ [c, d].
y=y
ZZ I
∂P
a) We firstly prove that − ∂y
dxdy = P dx. Indeed,
D ∂D
ZZ Z b ÃZ ψ(x) !
∂P ∂P
− dxdy = − (x, y)dy dx =
∂y a ϕ(x) ∂y
D
Z b Z b
= P (x, ϕ(x))dx − P (x, ψ(x))dx =
a a
I I I
= P dx + P dx = P dx.
arc[ACB] arc[BDA] ∂D
ZZ I
∂Q
b) We prove now that ∂x
dxdy = Qdx. Indeed,
D ∂D
ZZ Z Ã !
d Z β(y)
∂Q ∂Q
dxdy = (x, y)dx dy =
∂x c α(y) ∂x
D
246 6. DOUBLE INTEGRALS
Z d Z d
Q(β(y), y))dy − Q(α(y), y))dy =
c c
I I I
Qdy + Qdy = Qdy.
arc[CBD] arc[DAC] ∂D
We now add the two equalities obtained in a) and b) and we finally get
the Green formula. ¤
Remark 29. If our domain D is not simple with respect to one
axis or to both, we break it into a finite number of subdomains which
are simple w.r.t. both axis and they preserve the other properties of D.
Then we apply Green formula for each of these last domains. Let us
do this for the domain D of Fig.10,
Figure 10
I I
(3.6) + P dx + Qdy + P dx + Qdy+
arc[EF G] arc[GE]
I I
+ P dx + Qdy + P dx + Qdy.
arc[CDE] arc[EC]
I I I
Since P dx + Qdy + P dx + Qdy = 0 and P dx + Qdy +
arc[CE] arc[EC] arc[EG]
I
P dx + Qdy = 0, in formula (3.6) it remains
arc[GE]
ZZ µ ¶ I I
∂Q ∂P
− dxdy = P dx + Qdy + P dx + Qdy+
∂x ∂y
D arc[GHABC] arc[CDE]
I I
+ P dx + Qdy = P dx + Qdy,
arc[EF G] ∂D
Figure 11
3. GREEN FORMULA. APPLICATIONS. 249
yi − yi−1 1 ¯1
k+2 ¯
= · [xi−1 + t(xi − xi−1 )] ¯ =
xi − xi−1 k + 2 0
k+2 k+2
1 x − xi−1
= (yi − yi−1 ) i .
k+2 xi − xi−1
Hence
ZZ
(3.9) xk dxdy ≈
D
250 6. DOUBLE INTEGRALS
" #
1 xk+2 − xk+2 Xn
xk+2
i − xk+2
i−1
0 n
≈ (y0 − yn ) + (yi − yi−1 ) .
(k + 1)(k + 2) x0 − xn i=1
xi − xi−1
This last formula is important in practice. For instance, if D is bounded
by a polygonal line [A0 A1 ...An−1 An A0 ] and if k = 0, we get
" #
1 X n
(3.10) area(D) = (y0 − yn )(x0 + xn ) + (yi − yi−1 )(xi + xi−1 )
2 i=1
Here we do not need to put the absolute value for the determinant be-
cause the direct trigonometric direction of the sequence of points A0 ,
A1 , A2 assures the nonnegativity of this determinant. We propose to
the reader to find nice geometrical or mechanical interpretations for
formulas (3.9) and (3.10). Even if the global interpretation of the old
formula (4.5) is not clear enough.
We can even continue to find "nice" formulas. Let the m×m matrix
⎛ ⎞
ZZ
⎝Iij (D) = xi y j f (x, y)dxdy ⎠
D
Example 79. Let us use the Green formula (see theorem 80) to
I
compute the line integral I = xydx + x2 dy (see Fig.12). Thus,
arc[OABO]
ZZ Z 1 µZ 2x ¶ Z 1
1
I= (2x − x)dxdy = x dy dx = x2 dx = .
0 x 0 3
D
Figure 12
ZZ Z ÃZ √ !
3 9−x2
mass(D) = y 2 dxdy = x2 √
y 2 dy dx =
−3 − 9−x2
D
252 6. DOUBLE INTEGRALS
Z Z
2 3 ¯√9−x2 2 3 √
= 2
x y 3 ¯0 dx = x2 (9 − x2 ) 9 − x2 dx.
3 −3 3 −3
Here we must be careful because the same letter x is used for the
variable x and, at the same time, like a name of a function x(u, v) of
variables u and v. The same is for the letter y. We intentionally did
this because we wanted to show that the variable x was changed with
a function x(u, v) of variables u and v and, the variable y was changed
with a function y(u, v) of the new variables u and v. This change of the
"old" variables x and y with the "new" ones u and v is called a change
of variables. More exactly, the deformation − →
g : Ω → D is called a
change of variables if it is a diffeomorphism (see Fig.13)
Let us partially divide the domain Ω into rectangles Ωij = [ui−1 , ui ]×
[vj−1 , vj ], i = 1, 2, ..., n and j = 1, 2, ..., m. In each rectangle Ωij we
choose a fixed marked point Pij (ξ i , η j ). A small rectangle [P0 P1 P2 P3 ] =
[ui−1 , ui ] × [vj−1 , vj ] is deformed by −→g into a curvilinear parallelogram
[M0 M1 M2 M3 ] (see Fig.13 and Fig.14). First of all we want to estimate
the area of this last curvilinear parallelogram [M0 M1 M2 M3 ] in lan-
guage of the area of the rectangle [P0 P1 P2 P3 ] and of the deformation
functions x(u, v) and y(u, v).
¯ ¯
Theorem 81. area([M0 M1 M2 M3 ]) ≈ ¯J(ξ i , η j )¯ ·area([P0 P1 P2 P3 ]),
where J(ξ i , η j ) is the value of the Jacobian
µ ∂x ¶
∂u
(u, v) ∂x
∂v
(u, v)
J(u, v) = det ∂y
∂u
(u, v) ∂y
∂v
(u, v)
4. CHANGE OF VARIABLES IN DOUBLE INTEGRALS. POLAR COORDINATES.
253
Figure 13
of −
→
g (u, v) computed at the point Pij (ξ i , η j ). Here the sign "≈" means
"well approximation". This approximation is better and better if the
norms of the divisions generated by {ui } and {vj } are smaller and
smaller, i.e. if ui − ui−1 → 0 and vj − vj−1 → 0 for any i and j.
The above formula is usually written in a formal way:
(4.1) dxdy = |J(u, v)| dudv.
We say that the initial element of area dudv (which is the area of a
small rectangle!), after the deformation process generated by − →g , is "di-
lated" and becomes |J(u, v)| dudv. The dilation (or contraction) coeffi-
cient is exactly |J(u, v)| in a small neighboring of the point (u, v). This
means that the changes of areas are perfectly controlled by the Jacobian
matrix of the deformation − →g.
Proof. We approximate the area of the curvilinear parallelogram
[M0 M1 M2 M3 ] with the area of the parallelogram [M0 M1 M2 M3∗ ],
−−−−→ −−−−→
generated by the vectors M0 M1 and M0 M2 (see Fig.14). °−−−−→ −−−−→°
° °
But the area of this last parallelogram is equal to °M0 M1 × M0 M2 ° .
−−−−→ −−−−→
Let compute M0 M1 × M0 M2 . Firstly we have:
−−−−→ →
− →
−
M0 M1 = [x(ui , vj−1 ) − x(ui−1 , vj−1 )] i +[y(ui , vj−1 ) − y(ui−1 , vj−1 )] j =
∙ ¸
∂x → ∂y 0
− →
−
= (ci , vj−1 ) i + (c , vj−1 ) j (ui − ui−1 ).
∂u ∂u i
Here we used Lagrange formula for functions t à x(t, vj−1 ) and t Ã
y(t, vj−1 ) on the interval [ui−1 , ui ]. Since the length of this last interval
is small enough, since both functions just defined are of class C 1 , since
254 6. DOUBLE INTEGRALS
Figure 14
X
n X
m
¯ ¯
≈ f (x(ξ i , η j ), y(ξ i , η j )) · ¯J(ξ i , η j )¯ · (ui − ui−1 )(vj − vj−1 ) ≈
i=1 j=1
ZZ
≈ f (x(u, v), y(u, v)) |J(u, v)| dudv.
Ω
ZZ
Hence the two fixed numbers f (x, y)dxdy and
D
ZZ
f (x(u, v), y(u, v)) |J(u, v)| dudv become closer and closer! There-
Ω
fore they must coincide, i.e.
ZZ ZZ
f (x, y)dxdy = f (x(u, v), y(u, v)) |J(u, v)| dudv,
D Ω
what we wanted to prove. ¤
256 6. DOUBLE INTEGRALS
It is easy to see that in this last case J = abρ and our integral becomes
(we simply apply change of variables formula (4.4):
Z Z Z 1 µZ 2π ¶
2 2 2
x dxdy = ρ cos θdθ abρdρ =
0 0
x2 2
D: + y2 ≤1
a2 b
Z 1 µZ 2π ¶ µZ 1 ¶ µZ 2π ¶
3 2 3 2
= ab ρ cos θdθ dρ = ab ρ dρ cos θdθ =
0 0 0 0
Z 2π
1 1 + cos 2θ πab
= ab dθ = .
4 0 2 4
2 2
Example 82. The area of the ellipse xa2 + yb2 ≤ 1 can be immediately
computed by using the above change of variables.
Z Z Z 1 µZ 2π ¶
area = dxdy = dθ abρdρ = πab.
0 0
x2 2
D: + y2 ≤1
a2 b
Example 83. Sometimes the disc has not its centre in the origin of
the axes. For instance, let us compute the mass centrepof the disc D :
x2 + y 2 + 4y ≤ 0 if the density function is f (x, y) = x2 + y 2 . Since
the equation of the disc can be written as x2 + (y + 2)2 ≤ 4, we see
that the centre of the disc is in C(0, −2) and its radius is equal to 2. It
is symmetric w.r.t. Oy-axis, geometrically and also from the point of
view of statics (the mass is symmetric expanded w.r.t. Oy-axis). Hence
258 6. DOUBLE INTEGRALS
xG = 0. Let us compute
ZZ p
y x2 + y 2 dxdy
yG = ZDZ .
p
x2 + y 2 dxdy
D
To compute these two double integrals we shall use the common polar
coordinates (because the circle ∂D contains the origin! Otherwise this
method fails!). Let us write the Cartesian equation of the disc in the
language of the polar coordinates: ρ2 cos2 θ + ρ2 sin2 θ + 4ρ sin θ ≤ 0, or
ρ + 4 sin θ ≤ 0, or 0 < ρ ≤ −4 sin θ. This time, for any fixed value of θ,
ρ goes between 0 and −4 sin θ (Draw and look at it!). Now θ itself goes
between π and 2π. Hence
ZZ p Z 2π µZ −4 sin θ ¶ Z 2π 3 ¯−4 sin θ
ρ ¯¯
x2 + y 2 dxdy = ρ · ρdρ dθ = dθ =
π 0 3¯ π 0
D
Z 2π Z
64 64 2π3
=− sin θdθ = (1 − cos2 θ)d(cos θ) =
3 π 3 π
µ ¶¯2π
64 cos3 θ ¯¯ 256
= cos θ − ¯ = .
3 3 π 9
Now, we look at the last calculations and make some slight modifications
in order to compute the double integral in the numerator:
ZZ p Z 2π µZ −4 sin θ ¶
2
y x2 + y 2 dxdy = ρ sin θ · ρdρ dθ =
π 0
D
Z ï−4 sin θ ! Z
2π
ρ4 ¯¯ 256 2π 5
sin θ dθ = sin θdθ =
π 4 ¯0 4 π
Z 2π
= −64 (1 − cos2 θ)2 d(cos θ) =
π
Z 2π
= −64 (1 − 2 cos2 θ + cos4 θ)d(cos θ) =
π
∙ ¸¯2π
2 1 ¯ 16
−64 cos θ − cos θ + cos θ ¯¯ = −64 · .
3 5
3 5 π 15
−64· 16
Thus yG = 15
256 = − 12
5
.
9
4. CHANGE OF VARIABLES IN DOUBLE INTEGRALS. POLAR COORDINATES.
259
Example 84. Sometimes the centre of the disc is not in the origin
of the axes and the boundary of the disc does not passes through the
origin. For instance, let us compute the mass of the disc D : (x −
1)2 + (y + 3)2 ≤ 1 if the density function is f (x, y) = x. Let us use the
following change of variables:
½
x = 1 + ρ cos θ
, ρ ∈ (0, 1), θ ∈ [0, 2π).
y = −3 + ρ sin θ
Thus,
ZZ Z 1 µZ 2π ¶
mass(D) = xdxdy = (1 + ρ cos θ) dθ ρdρ = π.
0 0
D
Example 85. Let us compute for instance the mass of the plate
[A0 A1 A2 A3 ] with the density function f (x, y) = y (see Fig.15).
Figure 15
−−−→ →
− →
− −−−→ →
− →
−
Since A0 A1 = 10 i + 3 j and A2 A3 = 10 i + 3 j as free vectors,
we conclude that the domain D = [A0 A1 A2 A3 ] is a parallelogram. But,
ZZ
to compute the mass of D, i.e. the double integral ydxdy with the
D
iterative formulas, it is not so easy. Let us try to change this "com-
plicated" domain into a rectangle, on which it will be easy to integrate.
We need some knowledge from a Linear Algebra course. Let us start
with the square Ω = [0, 1] × [0, 1] and let us try to find a "translated"
linear mapping − →
g : R2 → R2 which "without the translation part"
is an isomorphism of vector spaces and such that − →
g (Ω) = D. Thus
260 6. DOUBLE INTEGRALS
−
→g (u, v) = (au +bv +e, cu + dv + h) (because of the linearity of the com-
ponent functions of − →g without the translation) and we only need to de-
termine the six numbers a, b, c, d, e, h and then to prove that − →
g (Ω) = D.
We even force that the corners of the square Ω = [0, 1] × [0, 1] to go
exactly to the "corresponding" corners of the parallelogram (our feeling
is that −→g is a dilation of axes + a translation + a rotation + a lin-
ear deformation of angles!). Hence, the conditions: − →g (0, 0) = (3, 2),
→
− →
− →
−
g (1, 0) = (13, 5), g (0, 1) = (4, 6) and g (1, 1) = (14, 9) (see Fig.6.15)
supply us with the values of a, b, c, d, e, h. In fact we do not need the
last equality because such a function − →g carries collinear vectors into
collinear vectors (Why?). Namely, a = 10, b = 1, c = 3, d = 4, e = 3
and hµ= 2. Thus →
−
¶ g (u, v) = (10u + v + 3, 3u + 4v + 2). Since the matrix
10 1
T = of the linear part (u, v) → (au+bv, cu+dv) of − →
g is in-
3 4
vertible, we see that − →g is a diffeomorphism from Ω to D, in particular
→
−g is a smooth deformation (without "defects") of Ω into D. Moreover,
it is also reversible, i.e. "one can come back in the same smoothly man-
ner"). The Jacobian of − →g is exactly the determinant of the matrix T.
Theorem 81 says that during the deformation process induced by − →
g the
→
−
area of Ω increases det T = 36 times. But why g (Ω) = D? Let us take
a point M ∈ [0, 1]×[0, 1] and let us write its position vector w.r.t. a fixed
−−→ →
− →
−
Cartesian system uO0 v : O0 M = α i0 + β j 0 where 0 ≤ α, β ≤ 1. We
need to prove that − →g (M) ∈ D, i.e. that (10α + β + 3, 3α + 4β + 2) ∈ D,
or, equivalently, that V = (10α + β, 3α + 4β), as a free vector in the
xOy-Cartesian system can be decomposed as a linear combination of
−−−→ →
− →
− −−−→ →
− →
−
vectors A0 A1 = 10 i + 3 j and A0 A2 = i + 4 j with subunitary
coefficients. Indeed, let us write:
→
− →
− →
− →
− →
− →
−
(10α + β) i + (3α + 4β) j = λ(10 i + 3 j ) + η( i + 4 j ),
or ½
10α + β = 10λ + η
.
3α + 4β = 3λ + 4η
Thus λ = α ∈ [0, 1] and η = β ∈ [0, 1], i.e. −
→
g (M) ∈ D. Now we are
ready to compute the mass of the plate D. The Jacobian is equal to 36,
thus
ZZ Z Z
mass(D) = ydxdy = (3u + 4v + 2) · 36dudv =
D [0,1]×[0,1]
Z 1 µZ 1 ¶ Z 1
= 36 (3u + 4v + 2)dv du = 36 (3u + 4) du =
0 0 0
4. CHANGE OF VARIABLES IN DOUBLE INTEGRALS. POLAR COORDINATES.
261
¶
µ
3
= 36 + 4 = 18 · 11 = 18 · (10 + 1) = 198.
2
Until this moment we considered only double integrals on bounded
domains. A similar theory can be done for double integrals on un-
bounded domains (see also the improper integrals of the first type).
For instance, let us compute again (we computed it with the help
of an integral with a parameter)
R ∞ −x2 and almost immediately the famous
Poisson’s integral I = 0 e dx.
Z ∞ Z ∞ Z Z
2 −x2 −y2 2 2
I = e dx · e dy = e−x −y dxdy =
0 0
D: x≥0,y≥0
Z ÃZ π
! µ ¶¯∞
x=ρ cos θ
∞
−ρ2
2 π 1 −ρ2 ¯¯ π
= ρe dθ dρ = − e ¯ = .
y=ρ sin θ 0 0 2 2 0 4
√
Z Z
π 2 2
Thus I = 2
. The integral e−x −y dxdy is an improper dou-
D: x≥0,y≥0
ble integral of the first type (the function is bounded but the domain is
unbounded).Formula (4.5) and the theory of improper simple integrals
are sufficient for the practical problems related to improper double in-
tegrals.
For instance, let us consider the following improper double integral
Z Z
√ 1
of the second type J = 3 2 2
dxdy. In this case the domain
x +y
D: x2 +y2 ≤1
1
is bounded but the function f (x, y) = √
3
is not bounded at (0, 0).
x2 +y2
Like in the situation of an improper simple integral of the second type,
we isolate the singularity by a neighborhood (a small disc in the 2-D
case) Dε : x2 + y 2 ≤ ε2 , for a small ε > 0 and write:
Z Z
1
J = lim p dxdy =
ε→0 3
x2 + y 2
DÂDε : ε2 ≤x2 +y2 ≤1
Z 1 µZ ¶ #¯1 "
ρ 3 ¯¯
2π 4
x=ρ cos θ − 23 3
= lim ρ ρdθ dρ = 2π lim 4 ¯ = π.
y=ρ sin θ ε→0 ε 0 ε→0
3
¯ 2
ε
Here we used the formula dxdy = ρdρdθ. Pay attention, do not forget
ρ, the Jacobian of the deformation −
→g , i.e. the dilation coefficient of
areas when one passes from rectangles to discs!
262 6. DOUBLE INTEGRALS
Triple integrals
Figure 1
Let now D = [a, b] × [c, d] × [e, g] be a parallelepiped space domain
and let us consider the following divisions ∆x : a = x0 < x1 < ... <
xn = b, ∆y : c = y0 < y1 < ... < ym = d, ∆z : e = z0 < z1 <
... < zp = g of the segments [a, b], [c, d] and [e, g] respectively. Let
∆ = ∆x × ∆y × ∆z = {Dijk = [xi−1 , xi ] × [yj−1 , yj ] × [zk−1 , zk ]} be the
3-dimensional division of D generated by ∆x , ∆y and ∆z (see Fig.2).
Let
{Pijk (ξ i , η j , θk )}, ξ i ∈ [xi−1 , xi ], η j ∈ [yj−1 , yj ], θk ∈ [zk−1 , zk ]
(i.e. Pijk ∈ Dijk ) be a fixed set of marking points of the division ∆ and
let
k∆k = max{diam(Dijk )}
i,j,k
be the norm of the division ∆. Let f : D → R be a density function
defined on D. If k∆k is small enough, we can well approximate the
mass of the solid (Dijk , f |Dijk ) by the number f (ξ i , η j , θk ) · vol(Dijk ),
i.e. with the density at the fixed point Pijk multiplied by the volume
1. WHAT IS A TRIPLE INTEGRAL ON A PARALLELEPIPED? 267
Figure 2
for any division ∆ with k∆k < δ ε and for any set of marking points
ZZZ
{Pijk (ξ i , η j , θk )} of ∆. This number I is denoted by f (x, y, z)dxdydz
D
and it is called the triple integral of f on the domain D. Here dxdydz is
called an element of volume and it symbolizes a small volume vol(Dijk ).
In general, dxdydz is not equal to the product dx · dy · dz. But here it
is so, because the volume of a small parallelepiped Dijk is "virtually"
268 7. TRIPLE INTEGRALS
ZZZ
equal to dx · dy · dz. We see that I = f (x, y, z)dxdydz is the mass
D
ZZZ
of the solid (D, f ) and vol(D) = dxdydz because for f = 1 each
D
Riemann sum Sf (∆; {Pijk (ξ i , η j , θk )}) is equal to vol(D).
It is easy to see that the above defined number I is unique.
Introducing Darboux sums like in the case of the double or simple
integrals, it is not difficult to prove a Darboux type criterion and the
following basic result.
Theorem 83. (iterative formulas for parallelepipeds) Let D = [a, b]×
[c, d] × [e, g] be a parallelepiped space domain and let f : D → R be a
piecewise continuous (see the definition above) function defined on D.
Then f is Riemann integrable on D and
ZZZ Z b µZ d µZ g ¶ ¶
(1.2) f (x, y, z)dxdydz = f (x, y, z)dz dy dx.
a c e
D
Moreover, the order of integration in this formula can be changed. This
formula says that first of all we
R g compute the simple integral with two
parameters x and y, J(x, y) = e f (x, y, z)dz, then we compute the sim-
Rd
ple integral K(x) = c J(x, y)dy with one parameter x, and finally we
ZZZ
Rb
compute a K(x)dx and find the value of f (x, y, z)dxdydz. This
D
means that the computation of a triple integral on a parallelepiped do-
main reduces to the calculation of three simple integrals, step by step,
i.e. in an iterative style.
Proof. For simplifying the reasoning, we assume that f is contin-
uous on D. Let us denote
Z b µZ d µZ g ¶ ¶
T = f (x, y, z)dz dy dx
a c e
and ZZZ
I= f (x, y, z)dxdydz.
D
Since, by definition, I exists if it is the unique limit point of the Rie-
mann sums Sf (∆; {Pijk (ξ i , η j , θk )}), we shall prove that T itself is the
unique limit point of such type of sums. So I exists and it will be equal
to T. Let us preserve the notation from the statement of the theorem.
1. WHAT IS A TRIPLE INTEGRAL ON A PARALLELEPIPED? 269
Rg
Since J(x, y) = e f (x, y, z)dz can be well approximated by Riemann
sums of the form:
p
X
J(x, y) ≈ f (x, y, θk )(zk − zk−1 ),
k=1
Rd
we see that K(x) = c
J(x, y)dy can be well approximated by sums of
the following type:
p µZ ¶
X d
(1.3) K(x) ≈ f (x, y, θk )dy (zk − zk−1 ).
k=1 c
Rd
Each integral c f (x, y, θk )dy can be well approximated by Riemann
sums of the type:
Z d mk
X (k) (k) (k)
f (x, y, θk )dy ≈ f (x, η j , θk )(yj − yj−1 ).
c j=1
(k) (k) (k)
We put together all the divisions ∆y : c = y0 < ... < ymk = d, k =
1, 2, ..., p and form a bigger division ∆y : c = y0 < y1 < ... < ym = d.
Hence
p
X Xm
K(x) ≈ f (x, η j , θk )(yj − yj−1 )(zk − zk−1 ).
k=1 j=1
Rb
Now, T = K(x)dx can be well approximated by sums of the type:
a
p
X Xm µZ b ¶
T ≈ f (x, η j , θk )dx (yj − yj−1 )(zk − zk−1 ).
k=1 j=1 a
But
Z b njk
X (jk) (jk)
f (x, η j , θk )dx ≈ f (ξ (jk) , η j , θk )(xi − xi−1 ).
a i=1
(jk) (jk) (jk)
By putting together all the divisions ∆x : a = x0 < ... < xnjk = b,
j = 1, 2, ..., m, k = 1, 2, ..., p we get that
p
X X
m X n
T ≈ f (ξ i , η j , θk )(xi − xi−1 )(yj − yj−1 )(zk − zk−1 ) =
k=1 j=1 i=1
p
X
n X
m X
= f (ξ i , η j , θk ) · vol(Dijk ),
i=1 j=1 k=1
because of the commutative and associative properties of the usual ad-
dition of real numbers. Hence T can be well approximated by Riemann
sums of the type (1.1), i.e. I = T. ¤
270 7. TRIPLE INTEGRALS
Z 1 µZ 1 µZ 1 ¶ ¶
= (3xyz + 1)dz dy dx =
0 0 0
Z 1 ÃZ ∙ ¸¯z=1 !
1
3xyz 2 ¯
= + z ¯¯ dy dx =
0 0 2 z=0
Z 1 µZ 1 ∙ ¸ ¶ Z 1∙ ¸¯y=1
3xy 3xy 2 ¯
+ 1 dy dx = + y ¯¯ dx =
0 0 2 0 4 y=0
Z 1∙ ¸ ∙ 2 ¸¯x=1
3x 3x ¯ 11
= + 1 dx = + x ¯¯ = .
0 4 8 x=0 8
Formulas for mass centres, static moments relative to coordinates
planes and inertia moments of solids can be immediately written by a
simple analogy with the similar formulas for laminas. Let (D, f ) be a
solid, where D is a parallelepiped space domain. Then the coordinates
of the mass centre G of D are:
ZZZ ZZZ
xf (x, y, z)dxdydz yf (x, y, z)dxdydz
(1.4) xG = Z ZDZ , yG = Z ZDZ ,
f (x, y, z)dxdydz f (x, y, z)dxdydz
D D
ZZZ
zf (x, y, z)dxdydz
zG = Z ZDZ .
f (x, y, z)dxdydz
D
If the solid (D, f ) is rotating around Oz-axis say, then the inertia mo-
ZZZ
ment Iz = (x2 + y 2 )f (x, y, z)dxdydz, etc. All of these formulas
D
can be deduced exactly like in the case of double integrals.
2. TRIPLE INTEGRALS ON A GENERAL DOMAIN. 271
Figure 3
The volume of such an elementary domain is equal, by definition,
to the sum of all volumes of the component parallelepipeds. Let us
recall definition 7 of a volume (measure in the 3-dimensional case).
Definition 27. A subset D of R3 has a volume vol(D) ∈ R+ if for
any small real number ε > 0 there exist two elementary domains Eint
and Eext such that:
1) Eint ⊂ D ⊂ Eext ,
2) vol(Eext ) − vol(Eint ) < ε.
3) vol(Eint ) ≤ vol(D) ≤ vol(Eext ).
This means that a subset D ⊂ R3 has a volume vol(D), a nonneg-
ative real number, if and only if there exist two towers of elementary
(j) (j)
domains {Eint } and {Eext } such that
(2.1)
(1) (2) (k) (m) (2) (1)
Eint ⊂ Eint ⊂ ... ⊂ Eint ⊂ ... ⊂ D ⊂ ... ⊂ Eext ⊂ ... ⊂ Eext ⊂ Eext ,
272 7. TRIPLE INTEGRALS
(m) (k)
vol(Eext ) & vol(D) and vol(Eint ) % vol(D) as m and k are convergent
(m) (m)
to ∞ (become larger and larger). Since vol(Eext ) − vol(Eint ) → 0 as
m → ∞, vol(D) is the unique real number which is contained in all
(m) (m)
intervals [vol(Eint ), vol(Eext )], m = 1, 2, .... (see Cantor’s axiom in
[Po]).
Using this definition, it is not difficult to prove that a finite set
of points, of segments of finite length smooth curves, or of smooth
surfaces, have volume zero. In particular, the boundary ∂D of a general
domain D has volume zero. As a consequence, a general domain D has
a finite volume, vol(D). It is also easy to see that a bounded subset A
of R3 , which has no interior points, has a volume vol(A) and this last
one is equal to zero.
The volume function D Ã vol(D) is an additive function, i.e.
vol(D1 ∪ D2 ) = vol(D1 ) +vol(D2 )−vol(D1 ∩D2 ) and it is an increasing
function, i.e. D ⊂ D0 implies vol(D) ≤ vol(D0 ).
Let now (D, f ) be a solid, i.e. D is a general domain and f :
D → R is a piecewise continuous function. A division (partition)
∆ of D is a finite set of general subdomains {Di }, i = 1, 2, ..., n of
D such that ∪ni=1 Di = D and for i 6= j, Di ∩ Dj is a subset with
no interior points (the empty set, a finite number of points, smooth
curves, smooth surfaces, etc.). The norm of ∆ is the nonnegative real
number
n°−−k∆k ° = max {diam(D o i ) : i = 1, 2, ..., n} , where diam(Di ) =
° −→0 ° 0
sup °MM ° : M, M ∈ Di . We usually denote ∆ = {Di }. A set of
marking points {Pi (ξ i , η i , θi )}, Pi ∈ Di , i = 1, 2, ..., n, is usually associ-
ated to a division ∆. Let
X
n
Sf (∆, {Pi (ξ i , η i , θi )}) = f (ξ i , η i , θi )vol(Di )
i=1
and
Mi = sup{f (x, y, z) : (x, y, z) ∈ Di } = f (x0(i) )
(see [Po], theorem 58). Hence
X
n
£ ¤
S∆ (f ) − s∆ (f ) = f (x0(i) ) − f (x(i) ) · vol(Di ) ≤
i=1
X
n
¯ ¯ ε X
n
≤ ¯f (x0(i) ) − f (x(i) )¯ · vol(Di ) < vol(Di ) = ε,
i=1
vol(D) i=1
° °
because k∆k < δ implies diam(Di ) < δ and so, °x0(i) − x(i) ° < δ.
Thus we can apply the uniform continuity of f and finally we get that
S∆ (f ) − s∆ (f ) < ε, i.e. that f is integrable. ¤
Remark 31. Up to now we used closed bounded domains Di with
piecewise smooth boundaries, as subdomains of a division ∆ of a fixed
domain D of the same type. Since any such subdomain Di can be well
approximated with elementary domains (finite union of nonoverlaping
parallelepipeds) or with finite unions of nonoverlaping tethrahedrons,
balls, prisms, etc., we can substitute these Di in definition 28 or in
theorem 85 with tethrahedrons, balls, prisms, etc. It is clear that in
general it is not possible to cover a domain D with nonoverlaping balls,
for instance. But, for any small ε > 0, there exists such a union of
balls Uε ⊂ D, such that vol(DÂUε ) < ε. Hence we can work with the
following alternative equivalent definition for an integrable function f.
We say that f : D → R is (Riemann) integrable on D if there ex-
ists a real number I such that for any small ε > 0 and η > 0 there
exists a union Uη = {Bi } P of nonoverlaping balls, such that Uη ⊂ D,
vol(DÂUη ) < η and |I − ni=1 f (ξ i , η i , θi )vol(Bi )| < ε for any set of
marking points {Pi (ξ i , η i , θi )}, Pi (ξ i , η i , θi ) ∈ Bi , i = 1, 2, ..., n. Simi-
larly we can work with tethrahedrons, prisms, general parallelepipeds,
etc., instead of balls. This is why if one has a tower of domains like in
the formula (2.1),
(1) (2) (k)
Eint ⊂ Eint ⊂ ... ⊂ Eint ⊂ ... ⊂ D,
(k)
such that vol(Eint ) % vol(D), then the approximation formulas can be
used:
ZZZ ZZ Z
(2.2) f (x, y, z)dxdydz = lim f (x, y, z)dxdydz,
k→∞
D (k)
Eint
276 7. TRIPLE INTEGRALS
or
ZZZ ZZ Z
(2.3) f (x, y, z)dxdydz ≈ f (x, y, z)dxdydz
D (k)
Eint
ZZZ ZZZ
=α f (x, y, z)dxdydz + β g(x, y, z)dxdydz.
D D
e)
ZZZ
f (x, y, z)dxdydz = f (x0 , y0 , z0 ) · vol(D),
D
then the set A of points (x, y, z) for which f (x, y, z) > 0 has the volume
equal to zero (the proof of this statement is more difficult).
f) if ∆ = {Di }, i = 1, 2, ..., n is a division of D, then
ZZZ Xn ZZZ
f (x, y, z)dxdydz = f (x, y, z)dxdydz.
D i=1 Di
If our solid (D, f ) is rotating around the Oz-axis for instance, the
inertia moment can be computed as follows:
ZZZ
Iz = (x2 + y 2 )f (x, y, z)dxdydz,
D
Figure 4
when (x, y) runs on a plane domain Ω contained in the xOy-plane (see
Fig.5). This means that D is a cylindrical domain delimited "up and
down" by the explicitly described surfaces z = ϕ(x, y) and z = ψ(x, y).
Figure 5
280 7. TRIPLE INTEGRALS
Figure 6
The first double integral is nothing else then 6 times the area of a disc
with radius 2, thus its value is equal to 24π. Let us use polar coordinates
change of variables to compute the second double integral:
Z Z Z 2 µZ 2π ¶
2 2 2
(x + y )dxdy = ρ · ρdθ dρ =
0 0
x2 +y2 ≤4
Z µZ ¶ ¯2
2
3
2π
ρ4 ¯¯
= ρ dθ dρ = 2π ¯ = 8π.
0 0 4 0
Hence,
3
mass(Ω) = 24π − · 8π = 12π.
2
We see that in the above iterative formula first of all one computes
a simple integral and then one computes a double integral. We shall
see in the following that in some cases it is possible to compute firstly
a double integral and then a simple one. Moreover, in the next result
the general domain D which appeared in the statement of theorem 88
needs not to be simple.
Theorem 89. (a general iterative formula II) Let Ω be a general
domain and let (Ω, f ) be a solid such that [e, g] = prOz (Ω), S = Ω ∩
{Z = z}, where z ∈ [e, g], is the section of Ω realized by the plane Z = z
and Ωz is the projection of S on the xOy-plane (see Fig.7). Then,
⎛ ⎞
ZZZ Z g ZZ
(3.2) f (x, y, z)dxdydz = ⎝ f (x, y, z)dxdy ⎠ dz.
e
Ω Ωz
3. ITERATIVE FORMULAS FOR A GENERAL SPACE DOMAIN 283
ZZ
Here, first of all we compute a double integral I(z) = f (x, y, z)dxdy
Ωz Rg
with a parameter z and then we calculate a simple integral e
I(z)dz
of the last computed function I(z).
Figure 7
Figure 8
Both iterative formulas (3.1) and (3.2) refers to the Oz-axis. We can
easily imagine the other four types of formulas which refer to Oy-axis
and to Ox-axis.
Example 89. For instance, let us find the coordinates of the mass
centre of the domain D : x2 + z 2 ≤ 2y, 0 ≤ y ≤ 8 and the density
function f (x, y, z) = y. The matter is about a circular paraboloidal
domain with Oy as an axe of symmetry. The symmetry here is also
a static symmetry because the mass is symmetrically distributed w.r.t.
3. ITERATIVE FORMULAS FOR A GENERAL SPACE DOMAIN 285
the Oy-axis. Indeed, f (−x, y, −z) = y = f (x, y, z), i.e. the density of
the point M(x, y, z) is the same as the density of its symmetric point
M 0 (−x, y, −z) w.r.t. Oy-axis. Thus, the mass centre G is on the Oy-
axis, i.e. xG = zG = 0. Let us compute
ZZZ
y · ydxdydz
D
yG = Z Z Z .
ydxdydz
D
ZZZ
Let us firstly compute I = ydxdydz, the mass of the solid. Let us
D
use the first iterative formula (3.1). The projection of the domain on
the zOx-plane is the disc: x2 + z 2 ≤ 16, with radius equal to 4. For
a fixed point P (x, 0, z) in this last disc, all the points of D which are
projecting in P are the points of the space segment
x2 + z 2
[M1 (x, , z), M2 (x, 8, z)].
2
Thus, ÃZ !
ZZZ Z Z 8
ydxdydz = ydy dxdz =
x2 +z 2
2
D x2 +z 2 ≤16
Z Z ∙ ¸ Z Z ∙ ¸
1 2 8 1 (x2 + z 2 )2
= y | x2 +z2 dxdz = 64 − dxdz =
2 2 2 4
x2 +z 2 ≤16 x2 +z 2 ≤16
Z Z Z Z
1
= 32 dxdz − (x2 + z 2 )2 dxdz.
8
x2 +z 2 ≤16 x2 +z 2 ≤16
The first integral is 32 times the area of a disc of radius 4, i.e. it is
equal to 32 × π × 16. Let us compute the second double integral
Z Z Z 4 µZ 2π ¶ ¯4
2 2 2 4 ρ6 ¯¯ 46
(x + z ) dxdz = ρ · ρdθ dρ = 2π ¯ = π .
0 0 6 0 3
x2 +z 2 ≤16
5 5
Hence, the mass of D is equal to 32 × π × 16 − π 46 = 43π .
Let us now compute the static moment w.r.t. xOz-plane,
ZZZ Z Z ÃZ !
8
2 2
J= y dxdydz = 2 2
y dy dxdz =
x +z
2
D x2 +z 2 ≤16
286 7. TRIPLE INTEGRALS
Z Z ∙ ¸ Z Z ∙ ¸
1 3 8 1 3 (x2 + z 2 )3
= y | x2 +z2 dxdz = 8 − dxdz =
3 2 3 8
x2 +z 2 ≤16 x2 +z 2 ≤16
Z Z Z Z
83 1
= dxdz − (x2 + z 2 )3 dxdz =
3 24
x2 +z 2 ≤16 x2 +z 2 ≤16
Z 4 µZ 2π ¶
83 1 6
= · 16π − ρ · ρdθ dρ =
3 24 0 0
¯4
83 1 ρ8 ¯¯ 83 1 48
= · 16π − · 2π · ¯ = · 16π − · 2π · =
3 24 8 0 3 24 8
1 ¡ 5 ¢
= 8 · 16π − 48 · 2π = 2π · 45 .
24 · 8
2π·45
Hence, yG = 45 π = 6.
3
Figure 9
The approximation means that the volume of the image of the paral-
lelepiped
[P0 P1 P4 P3 P2 P10 P40 P30 ]
through the deformation − →g is closer and closer to the quantity
¯ ¯
¯J−
→
g (ξ, η, θ) ¯ (u1 − u0 )(v1 − v0 )(w1 − w0 )
288 7. TRIPLE INTEGRALS
∂x
x2 − x0 ≈ (ξ, η, θ) · (w1 − w0 ),
∂w
∂y
y2 − y0 ≈ (ξ, η, θ) · (w1 − w0 ),
∂w
∂z
z2 − z0 ≈ (ξ, η, θ) · (w1 − w0 ),
∂w
∂x
x3 − x0 ≈ (ξ, η, θ) · (u1 − u0 ),
∂u
∂y
y3 − y0 ≈ (ξ, η, θ) · (u1 − u0 ),
∂u
∂z
z3 − z0 ≈ (ξ, η, θ) · (u1 − u0 ).
∂u
With these approximations, we come back to the above determinant
(4.5) and find:
⎛ ⎞
x1 − x0 y1 − y0 z1 − z0
det ⎝ x2 − x0 y2 − y0 z2 − z0 ⎠ ≈
x3 − x0 y3 − y0 z3 − z0
⎛ ∂x ∂y ∂z
⎞
∂v
(ξ, η, θ) ∂v
(ξ, η, θ) ∂v
(ξ, η, θ)
≈ det ⎝ ∂w ∂x
(ξ, η, θ) ∂w ∂y
(ξ, η, θ) ∂w ∂z
(ξ, η, θ) ⎠ (u1 −u0 )(v1 −v0 )(w1 −w0 ).
∂x ∂y ∂z
∂u
(ξ, η, θ) ∂u (ξ, η, θ) ∂u (ξ, η, θ)
Since the determinant of a transposed matrix is the same like the de-
terminant of the initial matrix and since by changing the rows between
them, the absolute value of the obtained determinant does not change,
we finally get that the absolute value of this last value is nothing else
then ¯ ¯
¯J−→ ¯
g (ξ, η, θ) (u1 − u0 )(v1 − v0 )(w1 − w0 ),
i.e. the formula which appeared in the statement of the theorem. Let
us recall the definition of the Jacobian of − →
g :
⎛ ∂x ∂x ∂x
⎞
∂u
(ξ, η, θ) ∂v (ξ, η, θ) ∂w (ξ, η, θ)
(4.7) J−
→g (ξ, η, θ) = det
⎝ ∂y (ξ, η, θ) ∂y (ξ, η, θ) ∂y (ξ, η, θ) ⎠
∂u ∂v ∂w
∂z ∂z ∂z
∂u
(ξ, η, θ) ∂v
(ξ, η, θ) ∂w
(ξ, η, θ)
¤
Let us use this basic result in order to prove the next theorem. We
shall again use the following simple observation: if two real numbers I
and J are limit points for one and the same subset A of real numbers
and if A has a unique limit point, then I = J. Instead of saying that I
is a limit point of A, we can say that I can be well approximated with
elements of A, etc.
290 7. TRIPLE INTEGRALS
Ω
4. CHANGE OF VARIABLES IN A TRIPLE INTEGRAL 291
Ω0
ZZZ
¯ ¯
= f (x(u, v, w), y(u, v, w), z(u, v, w)) · ¯J− ¯
g (u, v, w) dudvdw.
→
Ω
Here we used an immediate result (prove it!): "The integral (simple,
double, triple, etc.) on a subset of measure (length, area, volume, etc.)
zero is always equal to zero. Indeed, look carefully to any general Rie-
mann sum,...it is zero!
Remark 33. Moreover, during the proof of the above theorem 91
we did not used the fact that D or Ω are closed. Even some parts of
their boundaries ∂D or ∂Ω are missing, the formula continues to work.
This will be the situation in some future applications, when either the
Jacobian J− →
−
g is zero at some points of ∂Ω, or g is not a bijection or a
→
diffeomorphism around some points of Ω (see bellow the change of the
Cartesian coordinates x, y, z into the spherical coordinates ρ, ϕ, θ).
Example 90. Let D be the skew parallelepiped [OACBDA0 C 0 B 0 ]
−→ − → → −
− → −−→ → −
− →
generated by the vectors OA = i + 3 j + k , OB = −3 i + k and
−−→ − → − → →
−
OD = i + j + 4 k (see Fig.10).
Assume that D is loaded with a density function f (x, y, z) = x +
y + z. Let us compute the mass of D, i.e. the triple integral I =
292 7. TRIPLE INTEGRALS
Figure 10
ZZZ
f (x, y, z)dxdydz. We see that any of the iterative methods fails
D
in this case because of the "complicated" geometrical form of D (it can
also be divided into "many" simple subdomains w.r.t. Oz-axis). Let us
use the change of variables method described above. For this, we try
to find a linear transformation − →g : R3 → R3 such that the image of
the cube Ω = [0, 1] × [0, 1] × [0, 1] through − →
g be exactly our domain
D. Any course of Linear Algebra tells us that there exists one and only
one linear transformation (an isomorphism or automorphism in fact!)
which changes the basis {(1, 0, 0), (0, 1, 0), (0, 0, 1)} into another basis
−→ −−→ −−→
{OA, OB, OD}. Here we just identified the arithmetical vector space
R3 with the geometrical 3-D space V3 of all free space vectors. It is
easy to write the matrix M− →g of this linear transformation w.r.t. the
canonical basis {(1, 0, 0), (0, 1, 0), (0, 0, 1)} of R3 :
⎛ ⎞
1 −3 1
M− g =
→ ⎝ 3 0 1 ⎠.
1 1 4
Thus, −→g acts on the column vector (u, v, w)t ∈ R3 as follows:
⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞
u 1 −3 1 u u − 3v + w
→
−
g ⎝ v ⎠ = ⎝ 3 0 1 ⎠ ⎝ v ⎠ = ⎝ 3u + w ⎠ .
w 1 1 4 w u + v + 4w
Hence − →
g (u, v, w) = (u − 3v + w, 3u + w, u + v + 4w). Why − →
g (Ω) = D?
→
−
To answer to this question we can use the linearity of g . Indeed, an
element (u, v, w) ∈ R3 is in Ω = [0, 1] × [0, 1] × [0, 1] if and only if
4. CHANGE OF VARIABLES IN A TRIPLE INTEGRAL 293
Figure 11
= −ρ2 sin θ 6= 0,
the maximal definition domain of this regular transformation is: ρ ∈
(0, ∞), ϕ ∈ [0, 2π) and θ ∈ (0, π). We see that the maximal definition
domain of −→
g is the infinite parallelepiped Ω = [0, ∞) × [0, 2π] × [0, π]
without some of its exterior sides. The image of −→g is the whole space
R3 without the Oz-axis. During the triple integration process all of
these "gaps" of dimension 1 or 2 have zero volume, so that they means
"nothing" (see also remark 32). For instance, if we have to compute
4. CHANGE OF VARIABLES IN A TRIPLE INTEGRAL 295
ZZZ
the integral I = f (x, y, z)dxdydz, where D : x2 + y 2 + z 2 ≤ R2 is
D
the closed ball of radius R > 0 and with centre at O, we need in general
to use the change of variables formula by using spherical coordinates.
Let D∗ = DÂ{Oz-axis}, let ΩR = [0, R] × [0, 2π] × [0, π] and let Ω∗R =
(0, R] × [0, 2π) × (0, π). Then, using the remark 33, we get
ZZZ ZZZ
(4.11) f (x, y, z)dxdydz = f (x, y, z)dxdydz =
D D∗
ZZZ
(4.12) = f (ρ sin θ cos ϕ, ρ sin θ sin ϕ, ρ cos θ) · ρ2 sin θdρdϕdθ
Ω∗R
ZZZ
(4.13) = f (ρ sin θ cos ϕ, ρ sin θ sin ϕ, ρ cos θ) · ρ2 sin θdρdϕdθ.
ΩR
Hence,
ZZZ
(4.14) f (x, y, z)dxdydz =
D
ZZZ
= f (ρ sin θ cos ϕ, ρ sin θ sin ϕ, ρ cos θ) · ρ2 sin θdρdϕdθ,
ΩR
or
ZZ Z
(4.15) f (x, y, z)dxdydz =
D: x2 +y 2 +z 2 ≤R2
Z R ∙Z 2π µZ π ¶
¸
2
= ρ f (ρ sin θ cos ϕ, ρ sin θ sin ϕ, ρ cos θ) sin θdθ dϕ dρ.
0 0 0
Not always the too much processed formula (4.15) is useful. For in-
stance, let C(0, 0, a) be a point on the Oz-axis, a > 0 and let D be the
closed ball with centre at a and with radius a (see Fig.12).
Figure 12
ZZZ
Let us compute the inertia moment Iz = (x2 + y 2 )dxdydz of
D
the solid (D, 1) (or of D with the density function understood to be
f (x, y, z) = 1). Since the sphere ∂D contains the origin O, it is a
good idea to use the change of variables formula, by changing the old
Cartesian coordinates (x, y, z) with the spherical coordinates (ρ, ϕ, θ).
It is clear that if M(x, y, z) ∈ D, then the corresponding ϕ and θ verify
the obvious conditions: ϕ ∈ [0, 2π] and θ ∈ [0, π2 ] (see Fig.12). The
Cartesian relation which describes D is: x2 + y 2 + (z − a)2 ≤ a2 , or
4. CHANGE OF VARIABLES IN A TRIPLE INTEGRAL 297
"Z π
à ¯2a cos θ ! # Z π
2 ρ5 ¯¯ 26 a5 π 2
(4.16) = 2π sin3 θ dθ = sin3 θ cos5 θdθ.
0 5 ¯0 5 0
where r ∈ (0, ∞), ϕ ∈ [0, 2π) and z ∈ R. The slight modification in the
definition domain of the vector function − →
g (r, ϕ, z) = (r cos ϕ, r sin ϕ, z)
appeared because we want this function to be a smooth deformation
(here the Jacobian is equal to r, which must be not zero). As in the
case of spherical coordinates, what we got out (r = 0, i.e. the Oz-
axis, ϕ = 2π, i.e. the zOx-plane) are 1or 2-dimension pieces, so not
important during the triple integration process. The name "cylindrical
coordinates" comes from the fact that the parallelepiped Ω = [0, R] ×
[0, 2π] × [0, h] is deformed through the above cylindrical deformation − →g
2 2 2
into the cylindrical domain D : x + y ≤ R , 0 ≤ z ≤ h. This last
one is not "in fact’ a smooth deformation, but the exception points are
neglected during a process of triple integration, because a set of such
points has always zero volume. This is why we usually force things and
do not care of the true definition of a deformation!
where r ∈ [0, 4], ϕ ∈ [0, 2π] and x ∈ [3, 5]. Now we are really ready to
use the change of variables formula (4.8) with the Jacobian equal to r.
ZZZ ZZZ ZZZ
mass(D) = (2x + 1)dxdydz = 2 xdxdydz + dxdydz =
D D D
ZZZ ZZZ
=2 xdxdydz + vol(D) = 32π + 2 xdxdydz =
D D
Z 4 µZ 2π ∙Z 5 ¸ ¶
= 32π + r 2xdx dϕ dr =
0 0 3
Z µZ ¶ Z 4 µZ ¶
4 2π £ 2 5¤ 2π
= 32π + r x |3 dϕ dr = 32π + 16 r dϕ dr =
0 0 0 0
= 32π + 32π · 8 = 288π.
ZZZ
Let us compute now xf (x, y, z)dxdydz, i.e. the static moment
D
w.r.t. yOz-plane.
ZZZ ZZZ ZZZ
2
x(2x + 1)dxdydz = 2 x dxdydz + xdxdydz.
D D D
But the last triple integral on the right was just computed above, so
that:
ZZZ Z 4 µZ 2π ∙Z 5 ¸ ¶
2
x(2x + 1)dxdydz = 128π + 2 r x dx dϕ dr =
0 0 3
D
Z µZ 2π ¶
2 4 £ 3 3
¤ 3136 3520π
= 128π + r 5 − 3 dϕ dr = 128π + π= .
3 0 0 3 3
3520π
Finally, xG = 3·288π ≈ 4.074 ∈ [3, 5].
The next result is very important in Mechanics. It can be done for
simple, double, triple or surface integrals (these last ones will be studied
in the next chapter). We shall give it here only for triple integrals. We
leave as an exercise for the reader to state and to prove such a result
for simple, double or surface integrals.
Theorem 92. (the reduction theorem in Statics) Let
(D1 , f1 ), (D2 , f2 ), ..., (Dn , fn )
be n general nonoverlaping solids in R3 and let
G1 (x1 , y1 , z1 ), G2 (x2 , y2 , z2 ), ..., Gn (xn , yn , zn )
300 7. TRIPLE INTEGRALS
D1 Dn
mass(D1 )
· mass(D1 ) + ... + mass(Dn )
· mass(Dn )
= =
mass(D)
x1 mass(D1 ) + x2 mass(D2 ) + ... + xn mass(Dn )
= .
mass(D)
¤
4. CHANGE OF VARIABLES IN A TRIPLE INTEGRAL 301
Example 93. (a system of two solids) Let us apply this basic result
(theorem 92) to compute the coordinates xG , yG and zG , the coordinates
of the mass centre for the system D, formed with two homogenous solids
D1 , D2 (considered with the density function f = 1), where D1 is the
2 2 2
"north" hemisphere of the ellipsoid: xa2 + yb2 + zc2 ≤ 1, z ≥ 0 and D2 is
the "south" hemisphere of the ball: x2 +(y −2b)2 +z 2 ≤ b2 , z ≤ 0, where
a, b, c > 0. To compute the coordinates of the mass centres for D1 and
D2 respectively, we need appropriate changes of variables, inspired of
the usage of the spherical coordinates. For D1 we consider the "general
spherical coordinates", ρ, ϕ, θ :
⎧
⎨ x = x(ρ, ϕ, θ) = ρa sin θ cos ϕ,
D1 : y = y(ρ, ϕ, θ) = ρb sin θ sin ϕ ,
⎩ z = z(ρ, ϕ, θ) = ρc cos θ.
where ρ ∈ [0, 1], ϕ ∈ [0, 2π] and θ ∈ [0, π2 ]. The Jacobian of this trans-
formation is equal to −abcρ2 sin θ (prove this!). For instance, let us
compute the volume of D1 (which is equal to its mass because the den-
sity function is 1) by using formula 4.8:
ZZZ Z 1 "Z 2π ÃZ π ! #
2
vol(D1 ) = dxdydz = abc ρ2 sin θdθ dϕ dρ =
0 0 0
D1
µZ 1 ¶ µZ 2π ¶ ÃZ π
!
2
2 1 2πabc
= abc ρ dρ dϕ sin θdθ = abc · · 2π · 1 = .
0 0 0 3 3
2 2 2
Thus, the volume of the entire ellipsoid xa2 + yb2 + zc2 ≤ 1 is equal to 4πabc
3
.
4πR3
If a = b = c = R, we get the volume of a sphere of radius R : 3 . We
3
shall need bellow the volume of a half of a ball of radius b : vol = 2πb 3
.
ZZZ
zdxdydz
D1
centre of D1 . Since Oz is a symmetry axis for D1 , y1 = x1 = 0.
ZZZ Z 1 "Z 2π ÃZ π ! #
2
zdxdydz = abc2 ρ3 sin θ cos θdθ dϕ dρ =
0 0 0
D1
¯π
abc2 sin2 θ ¯¯ 2 πabc2
· 2π · = .
4 2 ¯0 4
302 7. TRIPLE INTEGRALS
πabc2
3c
Hence z1 = = 16
4
.
4πabc
3
Since D2 is symmetric w.r.t. the line x = 0, y = 2b, x2 = 0, y2 = 2b,
ZZZ
zdxdydz
D2
centre of D2 has the coordinates x2 , y2 and z2 . An appropriate change
of variables for D2 are:
⎧
⎨ x = x(ρ, ϕ, θ) = ρ sin θ cos ϕ,
y = y(ρ, ϕ, θ) = 2b + ρ sin θ sin ϕ ,
⎩ z = z(ρ, ϕ, θ) = ρ cos θ.
where ρ ∈ [0, b], ϕ ∈ [0, 2π] and θ ∈ [ π2 , π]. The Jacobian is obviously
J = −ρ2 sin θ, because the function y was modified by the addition of a
constant. Hence,
ZZZ Z b "Z 2π ÃZ π ! #
zdxdydz = ρ3 sin θ cos θdθ dϕ dρ =
π
0 0 2
D2
ÃZ !
π
b4 πb4
= sin θ cos θdθ · 2π · =− .
π
2
4 4
4
− πb4
Thus z2 = 2πb3
= − 3b
8
.
3
Let us use now formula (4.18) to compute xG , yG and zG .
0 · vol(D1 ) + 0 · vol(D2 )
xG = =0
vol(D)
2πb3
0 · vol(D1 ) + 2b · vol(D2 ) 2b ·
3 2b3
yG = = 3 =
vol(D) 2πabc
3
+ 2πb3
ac + b2
3c 3
16
· vol(D1 ) − 3b
8
· vol(D2 ) 3c 2πabc
16
· 3 − 3b 8
· 2πb
3 3 ac2 − b3
zG = = 3 = .
vol(D) 2πabc
3
+ 2πb
3
16 ac + b2
Many other practical properties of the mass centres can be proved
by simple applications of the basic properties of triple integrals. For
instance, for two nonoverlaping solids (D1 , f1 ), (D2 , f2 ) with their mass
centres G1 (x1 , y1 , z1 ) and G2 (x2 , y2 , z2 ), the mass centre G of the system
(D, f ), where D = D1 ∪ D2 , is on the segment which joins G1 and G2
(prove it!). What happens when G1 = G2 ?
5. PROBLEMS AND EXERCISES 303
0 ≤ z ≤ 1; f (x, y, z) = x2 + y 2 ;
c) x2 + y 2 ≤ 1, 1 ≤ z ≤ 2, f (x, y, z) = 2z;
4. Find the coordinates of the mass centre of the following domains,
bounded by:
a) x2 + y 2 = z, 0 ≤ z ≤ 1; b) x2 + y 2 = z, x + y + z = 0; c)
y 2 + 2z 2 = 4x, x = 2;
5. Find the inertia moment of the following solids w.r.t. Ox-axis:
a) x2 + y 2 + z 2 ≤ 9, f = 1; b) (x − 1)2 + y 2 + z 2 ≤ 1; c) x2 + y 2 ≤ 2x,
x ≥ 0, y ≥ 0, 0 ≤ z ≤ 2, f = 2z;
d) [0, 1] × [0, 2] × [0, 3]; e) x2 + z 2 ≤ y 2 , 0 ≤ y ≤ 2, f = 2y +
−→ −−→ −→
1; f) the general parallelepiped generated by OA, OB, OC, where
A(1, 3, 1), B(−2, 1, 1) and C(0, −1, 2); g) the tetrahedron [OABV ],
where A(−1, 3, 2), B(−3, 1, 4) and V (1, 1, 1).
6. Find the mass centre for D = D1 ∪D2 ∪D3 , where D1 : x2 +y 2 ≤ 9,
0 ≤ z ≤ 3, f1 (x, y, z) = z,
D2 : (x − 5)2 + y 2 + z 2 ≤ 4, f2 (x, y, z) = x and D3 : [10, 11] × [5, 6] ×
[0, 1], f3 (x, y, z) = z.
CHAPTER 8
Surface integrals
⎧
⎨ x = x(u, v)
(1.1) →
−
g : y = y(u, v) , (u, v) ∈ D .
⎩ z = z(u, v)
parametric representations:
⎧
⎨ x=x
→
−g1: y=y (x, y) ∈ R2 .
⎩ z = −2x + 3y + 1
C(u, v) →
−
+p k.
2 2 2
A (u, v) + B (u, v) + C (u, v)
(see Fig.1).
Figure 1
We are interested now how the area of a small rectangle [P0 P1 P2 P3 ],
where P0 (u0 , v0 ), P1 (u1 , v0 ), P2 (u0 , v1 ) and P3 (u1 , v1 ) changes when we
deform this rectangle through the deformation − →r (u, v) up to the sur-
face [M0 M1 M2 M3 ] (see Fig.1). The idea is the same like the idea used
in proving theorem 81.
Theorem 93. With the above notation and hypotheses, one has
(1.4) area[M0 M1 M2 M3 ] ≈
p
≈ A2 (u0 , v0 ) + B 2 (u0 , v0 ) + C 2 (u0 , v0 ) · area[P0 P1 P2 P3 ],
the approximation being better and better as the diameter of [P0 P1 P2 P3 ]
is smaller and smaller. Hence, the number
p
A2 (u0 , v0 ) + B 2 (u0 , v0 ) + C 2 (u0 , v0 )
acts as a dilation or contraction coefficient of the elementary area
area[P0 P1 P2 P3 ] = (u1 − u0 )(v1 − v0 ).
Formula (1.4) can also symbolically be written:
√
(1.5) dσ = A2 + B 2 + C 2 dudv,
1. DEFORMATION OF A 2D-DOMAIN INTO A SPACE SURFACE 309
Figure 2
this last deformation is said to be a cylindrical deformation. The
parametric equations of this cylinder are:
⎧
⎨ x = x(u, v) = R cos u
(1.8) →
−
g : y = y(u, v) = R sin u , (u, v) ∈ D = [0, 2π] × [0, h].
⎩ z = z(u, v) = v
1. DEFORMATION OF A 2D-DOMAIN INTO A SPACE SURFACE 311
Figure 3
2
The Cartesian equation of this last cone is x2 +y 2 = Rh2 z 2 , z ∈ [0, h].
√
Let us compute the expression A2 + B 2 + C 2 .
⎛ →
− →
− → ⎞
−
i j k
→
−
r u×− →r v = det ⎝ − Rh v sin u Rh v cos u 0 ⎠ =
R R
h
cos u h
sin u 1
R → R
− → R2 −
− →
= v cos u i + v sin u j − 2 v k .
h h h
312 8. SURFACE INTEGRALS
Thus
r r
√ R2 2 R4 2 R R2
A2 + B 2 + C 2 = v + 4v = v 1+ .
h2 h h h2
q
R R2
Hence dσ = h
v 1+ h2
dudv.
Figure 4
→
− →
− →
−
and let −
→r (ϕ, θ) = R sin θ cos ϕ i + R sin θ sin ϕ j + R cos θ k , ϕ ∈
[0, 2π] and θ ∈ [0, π] be a deformation of the rectangle D onto the sphere
x2 + y 2 + z 2 = R2 . In this case,
⎛ →
− →
− →
− ⎞
i j k
→
−r ϕ×− →r θ = det ⎝ −R sin θ sin ϕ R sin θ cos ϕ ⎠=
0
R cos θ cos ϕ R cos θ sin ϕ −R sin θ
→
− →
− →
−
= −R2 sin2 θ cos ϕ i − R2 sin2 θ sin ϕ j − R2 sin θ cos θ k .
√
Hence A2 + B 2 + C 2 = R2 sin θ.
In the case of a surface S which is explicitly given w.r.t. z, z =
z(x, y), the formulas can be putted in another more convenable form.
Indeed, the parametrization of S is:
x=x
y=y , (x, y) ∈ D.
z = z(x, y)
2. SURFACE INTEGRAL OF THE FIRST TYPE. THE MASS OF A 3D-LAMINA.
313
6)
ZZ
mass(S) = f (x, y, z)dσ,
S
where f is a density function on S.
7) The coordinates of the mass centre of the 3D-lamina (S, f ) are
calculated with the help of the following formulas:
(2.7)
ZZ ZZ ZZ
xf (x, y, z)dσ yf (x, y, z)dσ zf (x, y, z)dσ
xG = ZSZ , yG = ZSZ , zG = ZSZ .
f (x, y, z)dσ f (x, y, z)dσ f (x, y, z)dσ
S S S
ZZ p
= f (x(u, v), y(u, v), z(u, v)) A(u, v)2 + B(u, v)2 + C(u, v)2 dudv.
D
X
N
≈ f (x(ξ j , η j ), y(ξ j , η j ), z(ξ j , η j ))×
j=1
q
× A(ξ j , η j )2 + B(ξ j , η j )2 + C(ξ j , η j )2 · area(Rj ) ≈
2. SURFACE INTEGRAL OF THE FIRST TYPE. THE MASS OF A 3D-LAMINA.
317
ZZ p
≈ f (x(u, v), y(u, v), z(u, v)) A(u, v)2 + B(u, v)2 + C(u, v)2 dudv.
D
Hence,
ZZ
f (x, y, z)dσ ≈
S
ZZ p
≈ f (x(u, v), y(u, v), z(u, v)) A(u, v)2 + B(u, v)2 + C(u, v)2 dudv.
D
Since both these numbers are fixed numbers, they must coincide, i.e.
the equality (2.9) is completely proved. ¤
We can now compute different quantities which are expressed by
surface integrals of the first type.
Example 97. Let S : x2 + y 2 = 2z, 0 ≤ z ≤ 2 be a paraboloid (see
Fig.5).
Figure 5
Z 2 p µZ 2π ¶
x=ρ cos θ
= ρ ρ2 + 1 dθ dρ =
x=ρ sin θ 0 0
Z ¯
3 2
¡ 2 ¢ 12 2 ¯
2 2
(ρ + 1) ¯ 2π ¡ 3/2 ¢
=π ρ + 1 d(ρ2 + 1) = π 3 ¯ = 5 − 1 .
0 2
¯ 3
0
Z 2 p Z √
5
3 3 ρ2 +1=t2
= · 2π 2
ρ ρ + 1dρ = 3π t2 (t2 − 1)dt = etc.
2 0 1
Z p µZ ¶
3 2
3
2π £ 2 2 4
¤
= 2
ρ ρ +1 4ρ sin θ + ρ dθ dρ =
8 0 0
Z 2 p µZ 2π ¶ Z 2 p
3 5 2 3π
(2.12) = 2
ρ ρ +1 sin θdθ dρ + ρ7 ρ2 + 1dρ.
2 0 0 4 0
But
Z 2π Z 2π
2 1 − cos 2θ
sin θdθ = dθ = π,
0 0 2
so that 2.12 becomes
Z Z
3π 2 5 p 2 3π 2 7 p 2
Ix = ρ ρ + 1dρ + ρ ρ + 1dρ = etc.
2 0 4 0
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 319
Figure 6
(called the transition matrix from the basis {v1 , v2 , v3 } to the basis
{w1 , w2 , w3 }) has its determinant det C a positive real number. The
infinite set of all the bases in V3 , the real vector space of all 3D-free
vectors, can be divided into two nonoverlaping (their intersection is
empty) subsets A and A− . Two bases {v1 , v2 , v3 } and {w1 , w2 , w3 } be-
longs to the same subset if the transition matrix from one to another
has a positive determinant. Let denote by A the subset (or class) which
→ −
− → − →
contains the basis { i , j , k } and by A− the subset (or class) which
→ −
− → −→
contains the basis { i , j , −k}. We leave as an exercise for the reader
to prove that A ∪ A− contains all the bases of V3 and that A ∩ A− = ∅,
the empty set. We call A the direct orientation of the space V3 and
A− the inverse orientation of V3 . The same can be done for V2 , the real
vector space of the 2D-free vector space (i.e. the plane free vectors).
Thus, whenever we fix a Cartesian coordinate system uov in a plane
(P ), we just fixed a direct orientation:"from u to v". What is this?
→
−
The direction of the ou-axis is given by a versor i0 . The direction of
→
−
ov-axis is given by a versor j 0 . By definition the class A of orientation
→ −
− →
into which the base { i0 , j 0 } belongs is said to be the direct orientation
of the Cartesian system uov. "From u to v" means that we take firstly
→0
− →
−
i then, secondly, we take j 0 . We simply say that uov is an oriented
plane Cartesian coordinate system (here the direct orientation is "from
u to v” and the inverse one is "from v to u") (see Fig.7).The closed
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 321
−−−→
if −
→
n (M0 ) = OM0∗ ) on the unit sphere (US) : x2 + y 2 + z 2 = 1 in the
Oxyz-space (the space into which (S) is considered). Let us leave M0
to run freely on (S) and let us denote it by M. We obtain in this way
a new function M Ã M ∗ , which associates to a point M of (S), a
point M ∗ on (US). Let us call this function the orientation function
and denote it by Or : (S) → (US), Or(M) = M ∗ .
Figure 7
Figure 8
→ −
− → − →
It is easy to see that −→
n (M) = − a√ia+b j +c k
2 +b2 +c2 , i.e. the normal versor of the
B(u, v) →
−
+p j+
A(u, v)2 + B(u, v)2 + C(u, v)2
C(u, v) →
−
+p k,
A(u, v)2 + B(u, v)2 + C(u, v)2
1
then cos γ = √A2 +B 2 +1 > 0 i.e. γ ∈ [0, π2 ). Many times, this last
information is sufficient to determine the direct face (which corresponds
to the direct orientation) of the surface (S).
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 325
⎧
x=x
⎨
Example 99. Let (S) : yp= y , x2 + y 2 ≤ R2 be the
⎩ 2 2
z = x +y
conic surface of Fig.9.
Figure 9
We are interested to find the mathematical description of the normal
versors −
→n which define the exterior side of this conic surface (see −→
n
→
− ◦
in Fig.9). The angle γ between n and Oz is greater then 90 , so that
cos γ < 0, i.e.
1
cos γ = − p .
A(u, v) + B(u, v)2 + 1
2
Hence, we must put a minus in front of cos α, cos β and cos γ, in the
expression of −
→
n (see (3.2)). Therefore, in our case, for any M(x, y, z) ∈
(S), Ã !
→
− x y 1
n (M) = √ p ,√ p , −√ .
2 x2 + y 2 2 x2 + y 2 2
Remark 34. It is not difficult to see that if D is a smooth domain
(closed, bounded, connected and ∂D is a smooth curve) and if −→r :D→
→
−r (D) = (S) is a smooth deformation of D onto the surface (S) (− →
r is
1 →
− →
−
of class C , injective and r u × r v 6= 0 on D), then (S) is orientable.
Let now D be a closed, bounded and connected subset of the uov-
plane with its boundary ∂D a piecewise directed oriented curve in uov.
Let D = ∪ni=1 Di be a division of D with nonoverlaping (Di ∩ Dj has a
326 8. SURFACE INTEGRALS
zero area for i 6= j) oriented subdomains of the same type like D (see
Fig.10). We see in this figure that the intersection curve Di ∩ Dj (if it
exist like a curve!) is always double oriented. Moreover, the orientation
of each Di (i.e. the orientation of ∂Di ) is completely determined by
the orientation of D (i.e. of ∂D).
Figure 10
→
− →
− → →
−
Let − →r (u, v) = x(u, v) i + y(u, v) j + z(u, v) k , − r :D→−
→
g (D) =
(S) be a piecewise smooth deformation and let Pi (ui , vi ) ∈ Di , i =
1, 2, ..., n be a set of marking points of the division {Di } of D. Then
{Si = − →g (Di )}, i = 1, 2, ..., n is a division of (S). If
k{Di }k = max{diam(Di ) : i = 1, 2, ..., n}
becomes smaller and smaller, then k{Si }k becomes smaller and smaller,
except maybe a subset of points of zero area of D. Let Mi = − →
r (Pi ),
i.e. Mi (x(ui , vi ), y(ui , vi ), z(ui , vi )) for i = 1, 2, ..., n. We assume that
(S) and, as a consequence, all (Si ) are orientable surfaces. Let
r u (ui , vi ) × −
→
− →
r v (ui , vi )
→
−n (Mi ) = − =
→ →
−
k r (u , v ) × r (u , v )k
u i i v i i
A(ui , vi ) →
−
(3.3) =p i+
2 2
A(ui , vi ) + B(ui , vi ) + C(ui , vi )2
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 327
B(ui , vi ) →
−
+p j+
2 2
A(ui , vi ) + B(ui , vi ) + C(ui , vi )2
C(ui , vi ) →
−
+p k,
2 2
A(ui , vi ) + B(ui , vi ) + C(ui , vi )2
→
−
Definition 31. F is integrable on the direct oriented surface (S)
if there exists a real number I such that for any small real number
ε > 0, one can find another small real number δ > 0 (which de-
pends¯ on ε) such that if {D¯i } is a division of D with k{Di }k < δ,
¯ ¯
then ¯I − S−→
F ,(S)
({Si }, {Mi })¯ < ε for any set of marking points {Pi },
→
−
Pi ∈ Di . I is called the surface integral IofIthe second type of F on
→ −
−
the oriented surface (S). It is denoted by F ·→n dσ. The sign "+"
(S)+
328 8. SURFACE INTEGRALS
3) I I I I I I
− −
→ − −
→ − −
→
F ·→
n dσ = F ·→
n dσ + F ·→
n dσ,
(S)∪(T ) (S) (T )
where (T ) is another oriented surface such that on the common curve
∂(S) ∩ ∂(T ) the orientation induced by the orientation of (T ) must
be the inverse orientation of the same intersection curve viewed with
the orientation induced by the orientation of (S) (see the situation of
two neighboring oriented subdomains in Fig.10 or in Fig.13). If the
surfaces (S) and (T ) have no small smooth curve in common, then
the orientation of (T ) cannot be connected with the orientation of (S)
by anything else, so that in the union (S) ∪ (T ) they keep their own
orientation.
Let us now try to proveI the
I existence and to compute the surface
→ −
−
integral of the second type F ·→n dσ. For this we approximate the
(S)
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 329
p ZZ
= A(u∗ , v ∗ )2 + B(u∗ , v ∗ )2 + C(u∗ , v ∗ )2 dudv ≈
Di
p
≈ A(ui , vi )2 + B(ui , vi )2 + C(ui , vi )2 area(Di ),
because both points Pi (ui , vi ) and Pi∗ (u∗ , v ∗ ) are in the same small
domain Di . Let us come back in formula (3.4) with this expression of
the area(Si ) and with the expression of − →n (Mi ) from formula (3.3):
X h−
n
→ i
F (Mi ) · −→
n (Mi ) · area(Si ) ≈
i=1
X
n
≈ [P (Mi )A(Mi ) + Q(Mi )B(Mi ) + R(Mi )C(Mi )] area(Di ).
i=1
But this last sum is a Riemann sum of the double integral
ZZ
[P (x(u, v), y(u, v), z(u, v))A(u, v)+Q(x(u, v), y(u, v), z(u, v))B(u, v)+
D
(3.6)
ZZ
= [P (x(u, v), y(u, v), z(u, v))A(u, v)+Q(x(u, v), y(u, v), z(u, v))B(u, v)+
D
where cos α, cos β and cos γ are the coordinates of the normal versor
→
−n (M) at the point M(x, y, z). This is not a practical formula. But
it
I Iis useful in order to find a new expression for the surface integral
→ −
−
F ·→n dσ. If we look at Fig.11
(S)
Figure 11
we see that the area-projection prxOy dσ of an orientated small ele-
ment dσ (for instance the area of an (Si ))of the surface (S) on the xOy-
plane is a direct oriented area in this plane, usually denoted by dxdy. It
is clear enough that the order dx then dy in the symbol dxdy is essential,
because it reflects the direct orientation "from x to y" in the xOy-plane.
Hence dxdy = −dydx (dxdy is the "up" area and dydx is the "down"
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 331
this last being a formal expression, not used in practice. The expression
ω = P dydz + Qdzdx + Rdxdy is said to be a differential form of order
2 in three variables. Do not forget the order: dzdx and not dxdz!
I I
Example 100. Let us compute I = xdydz + ydzdx, where (S)
(S)
is the interior face of the paraboloid: z = x2 + y 2 , 0 ≤ z ≤ 4 (see
Fig.12).
Figure 12
∂z ∂z
we can use formula (3.2) and find A = − ∂x = −2x, B = − ∂y = −2y,
π
C = 1, γ ∈ [0, 2 ],
à !
→
− −2x −2y 1
n (x, y, z) = p ,p ,p .
1 + 4x2 + 4y 2 1 + 4x2 + 4y 2 1 + 4x2 + 4y 2
Let us use now the basic formula (3.6):
I I Z Z
xdydz + ydzdx = −2 (x2 + y 2 )dxdy =
(S) D:x2 +y 2 ≤4
Z 2 µZ 2π ¶ Z 2
3
= −2 ρ dθ dρ = −4π ρ3 dρ = −16π.
0 0 0
1
Remark 35. If (S) is not of class C but it is piecewisely of class
1
C (see Fig.13), then we give an orientation to each smooth pieces
of (S) such that on the common intersection curves, these last curves
must have 2 distinct type of orientation (each of them coming from
the orientation of the neighboring subdomains). In fact, the orienta-
tion of ∂(S) completely determines the orientation of all the component
subdomains.
Figure 13
→
−
Example 101. Let us compute the flux of the field F = (0, 0, z)
on the exterior faces of the tetrahedron [OABC] of Fig.14. Since the
surface is a union of 4 distinct surfaces with different parametrizations,
3. FLUX OF A VECTOR FIELD THROUGH AN ORIENTED SURFACE. 333
we successively compute:
I I
I4 = (0, 0, z) · −
→
n4 dσ.
[ABC]
Since
Figure 14
⎧
⎨ x=x
[ABC] : y=y , x + y ≤ 1,
⎩ z =1−x−y
A = 1, , B = 1, C = 1, so
Z Z Z 1 µZ 1−x ¶
I0 = (1 − x − y)dxdy = (1 − x − y)dy dx =
0 0
x+y≤1
Z µ ¶¯1−x Z 1µ 2 ¶
1
y 2 ¯¯ x 1 1
y − xy − ¯ dx = −x+ dx = .
0 2 0 0 2 2 6
Now I I
I1 = (0, 0, z) · −
→
n1 dσ = 0,
[OAB]
−
→
because −
→
n1 = − k and z = 0 on [OAB].
I I I I
→
−
I2 = (0, 0, z) · →
−
n2 dσ = (0, 0, z) · (− j )dσ = 0.
[OCA] [OCA]
334 8. SURFACE INTEGRALS
I I
→
−
I3 = (0, 0, z) · (− i )dσ = 0.
[OBC]
Figure 15
Figure 16
[XY ]⊥[Z 0 Z 00 ], we see that [XY ] is perpendicular on the entire plane
[ZZ 0 Z 00 ]. In particular it is perpendicular on [ZZ 00 ]. Let − →
n (Z 00 ) be the
normal versor on the plane [XY Z] at the point Z 00 (see Fig.16). Since
γ + δ = 90◦ and γ 0 + δ = 90◦ one has that γ = γ 0 (see Fig.16). Thus,
[Z 0 Z 00 ] = [ZZ 00 ] cos γ and so,
1 1 1
area(∆XY Z) = [XY ][ZZ 00 ] = [XY ] cos γ[ZZ 00 ] =
2 cos γ 2
1 1 1
= [XY ][Z 0 Z 00 ] = area(∆XY Z 0 ].
cos γ 2 cos γ
Hence, area(prxOy ∆[XY Z]) = area(∆XY Z) cos γ, what we wanted to
prove.
We must be very careful with the notation dxdy, the area of the
direct oriented domain which is the projection of a small part of the
direct oriented surface S, of area dσ. Since dxdy = −dydx, one uses the
exterior product notation dxdy = dx ∧ dy. To change dxdy with dydx
means to take the other face of the oriented projection ω (see Fig.17).
4. Problems and exercises
ZZ
1. Compute I = xydσ, where S is the part of the plane (P ) :
S
2x + y + 2z = 1 which is inside the cylinder: x2 + y 2 = 1.
2. The cylindrical solid x2 + y 2 ≤ R2 cut the semisphere x2 + y 2 +
z = 4R2 , z ≥ 0 into a surface S. Find its area and its centre of mass
2
Figure 17
ZZ
3. Compute (x2 + y 2 )dσ is S is a surface with the following
S
parametrization: x = u, y = v, z = uv and u2 + v2 ≤ 1.
4. Find the moment of inertia of the cone y 2 + z 2 = 4x2 , x ∈ [−2, 2]
with respect to the Oz-axis.
5. Find the centre of mass for the 3D-lamina (S, f ), where S is the
surface x2 + z 2 = 4y, y ≤ 4 and f (x, y, z) = y.
ZZ
6. Compute I = (x + y + z)dσ, where S is the union of all the
S
sides of the parallelepiped [0, 1] × [0, 2] × [0, 3].
ZZ
7. Compute I = (x + y + z)dσ, where S is the surface x2 + y 2 +
S
z 2 = a2 , z ≥ 0.
8. Find the mass of the 3D-plate (S, f ), where S is the surface
x2 + y 2 + z 2 = 1, x ≥ 0, y ≥ 0, z ≥ 0 and the density function is
−1
f (x, y, z) = [x2 + y 2 + (z − 1)2 ] 2 .
9. Find the area of the part of the sphere x2 + y 2 + z 2 = a2 , which
is inside the cylinder xI2 + 2
I y = ax.
10. Compute I = xydydz + xzdzdx + 2x2 ydxdy, where (S) is
(S)
the exterior side of the surface z = x2 + y 2 , x2 + y 2 ≤ 4.
→
−
11. Compute the flux of the field F (x2 , y 2 , z 2 ) through the interior
side of the surface x2 + y 2 = z, z ≤ 1.
4. PROBLEMS AND EXERCISES 337
→
−
12. Compute the flux of the field F (x3 , y 3 , z 3 ) through the total
exterior sides of the surface x2 + y 2 = z 2 , 0 ≤ z ≤ 4.
→
− →
− →
− →
−
13. Find the flux of the field F (x, y, z) = x i + y j + z k through
the interior side of the sphere x2 + y 2 + z 2 = R2 .
→
−
14. Find the flux of the field F (x2 , y 2 , z 2 ) through the total exterior
sides of tetrahedron [OABC], where A(1,³0, 2), B(0, 2, 0) and C(0, 0, 3).
→ 1 1 1´
−
15. Compute the flux of the field F x , y , z through the exterior
side of the sphere x2 + y 2 + z 2 = R2 .
→
−
16. Find the flux of the field F (yz, xz, xy) through the total exte-
rior face of the cylinder x2 + y 2 = a2 , 0 ≤ z ≤ h.
CHAPTER 9
−
→ →
− →
− →
−
F (x, y, z) = P (x, y, z) i + Q(x, y, z) j + R(x, y, z) k
Figure 1
Before to prove this basic result, let us make some commentaries
on it.
The philosophical importance of Gauss formula (1.1) is the follow-
ing. On the left side of this formula we have an "oriented" object, in
general very complicated, but only a 2D-object. On the right side one
has a triple integral, a simpler object (because it has no "orientation" at
all), but a 3D-object. Hence this formula connects a 2D-computation
with a 3D-computation. On the left side of this formula we have in
→
− →
−
fact the flux ΦS ( F ) of the field F through the surface S, "from inside
to outside", i.e. the normal versor at a point M ∈ S is "above" the
tangent plane at M to S (see Fig.1). On the right side of the formula
→
−
one has "the global productivity" of the volume Ω w.r.t. the field F .
Let us explain in the following the word "productivity". Let us apply
the mean formula (see theorem 87, e)) to the triple integral from the
right side:
ZZZ
→
− →
−
div F dxdydz = div F (P0 ) · vol(Ω),
Ω
1. GAUSS FORMULA (DIVERGENCE THEOREM) 341
Gauss succeeded to express this last law as a "local law" (at a point
P0 6= O!):
→
− ρ(P0 )
(1.3) div E (P0 ) = ,
ε0
where ρ(P0 ) is the charge density at P0 produced by q0 . It appears that
from here we can deduce our Gauss formula (1.1) in this particular case.
There is a mathematical problem! The electric field is not defined at
the origin O! Thus, our Gauss formula (1.1) cannot be applied in this
342 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
Figure 2
The triple integral can be naturally decomposed into three triple
integrals and so we can apply Gauss formula for the three simple do-
mains:
⎡ ⎤ ⎡ ⎤
ZZZ ZZZ ZZZ
→
− →
− →
−
div F dxdydz = ⎣ div F dxdydz ⎦+⎣ div F dxdydz ⎦ +
Ω Ω1 Ω2
⎡ ⎤
ZZZ
→
−
+⎣ div F dxdydz ⎦ =
Ω3
⎡ ⎤
I I I I I I
⎢ − −
→ − −
→ → − −
→ → ⎥
=⎣ F ·→
n1 dσ + F · n01 dσ + F · n001 dσ ⎦ +
(S)Â(∂Ω2 )Â(∂Ω3 ) (S1 ) (S2 )
⎡ ⎤
I I I I
⎢ − −
→ − −
→ → ⎥
+⎣ F ·→
n2 dσ + F · n02 dσ ⎦ +
(∂Ω2 )Â(S1 ) (S1 )
344 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
⎡ ⎤
I I I I
⎢ − −
→ − −
→ → ⎥
+⎣ F ·→
n3 dσ + F · n03 dσ ⎦ =
(∂Ω3 )Â(S2 ) (S2 )
⎡ ⎤
I I I I I I
⎢ − −
→ − −
→ − −
→ ⎥
=⎣ F ·→
n1 dσ + F ·→
n2 dσ + F ·→
n3 dσ ⎦ +
(S)Â(∂Ω2 )Â(∂Ω3 ) (∂Ω2 )Â(S1 ) (∂Ω3 )Â(S2 )
⎡ ⎤ ⎡ ⎤
I I I I I I I I
⎢ − −
→ → − −
→ → ⎥ ⎢ − −
→ → − −
→ → ⎥
+⎣ F · n01 dσ − F · n01 dσ ⎦+⎣ F · n001 dσ − F · n001 dσ ⎦ =
(S1 ) (S1 ) (S2 ) (S2 )
I I
− −
→
= F ·→
n dσ.
(S)
I I ZZZ
∂Q
(1.6) b) Q(x, y, z)dzdx = dxdydz,
∂y
(S) Ω
I I ZZZ
∂R
(1.7) c) R(x, y, z)dxdy = dxdydz.
∂z
(S) Ω
Since the proofs for all these formulae a), b) or c) are similar, we shall
prove only formula c). For this, after dividing the domain Ω into a finite
union of nonoverlaping subdomains (see what we did above!) which are
simple w.r.t. Oz-axis, we suppose that Ω is a simple domain w.r.t. Oz-
axis like in Fig.3. This means that
Ω = {(x, y, z) : ϕ(x, y) ≤ z ≤ ψ(x, y), (x, y) ∈ Ωxy = prxOy Ω} ,
where ϕ and ψ are continuous functions defined on Ωxy .
Let us look now at Fig.3 and write:
⎛ ⎞
ZZZ ZZ Z
ψ(x,y)
∂R ⎜ ∂R ⎟
dxdydz = ⎝ dz ⎠ dxdy =
∂z ∂z
Ω Ωxy ϕ(x,y)
1. GAUSS FORMULA (DIVERGENCE THEOREM) 345
Figure 3
ZZ ZZ
= R(x, y, ψ(x, y))dxdy − R(x, y, ϕ(x, y))dxdy =
Ωxy Ωxy
I I I I
= R(x, y, z)dxdy + R(x, y, z)dxdy+
(S2 ) (S1 )
I I I I
+ R(x, y, z)dxdy = R(x, y, z)dxdy,
(S3 ) (S)
I I
because R(x, y, z)dxdy = 0, (γ = 90◦ , so cos γ = 0 and dxdy =
(S3 )
Figure 4
→
−
Use then this result to compute the flux Φ1 of the same field F
through the exterior surface of the cube, except the hatched face [O0 A0 B 0 C 0 ].
I I ZZZ
→
− → −
− → Gauss
ΦS ( F ) = F · n dσ = (2x + 2y + 2z)dxdydz =
S Ω
⎛ ⎛ ⎞ ⎞
Z1 Z1 Z1
= 2 ⎝ ⎝ (x + y + z)dx⎠ dy ⎠ dz =
0 0 0
⎛ ⎞
Z1 Z1 à µ 2 ¶¯1 !
x ¯
=2 ⎝ + xy + xz ¯¯ dy ⎠ dz =
2 0
0 0
⎛ ⎞
Z1 Z1 µ ¶ Z1 à µ ¶¯1 !
1 y y 2 ¯
=2 ⎝ + y + z dy ⎠ dz = 2 + + yz ¯¯ dz =
2 2 2 0
0 0 0
Z1 µ ¶¯1
z 2 ¯¯
= 2 (1 + z) dz = 2 z + = 3.
2 ¯0
0
Let (S1 ) be the hatched surface [O0 A0 B 0 C 0 ]. Since the normal versor
→
− →
−
n = k = (0, 0, 1), one has that dydz = cos αdσ = 0 and dzdx =
cos βdσ = 0. Thus,
Z Z
→
−
Φ(S1 ) ( F ) = dxdy = 1.
[0,1]×[0,1]
1. GAUSS FORMULA (DIVERGENCE THEOREM) 347
→
− →
−
Hence Φ1 = ΦS ( F ) − Φ(S1 ) ( F ) = 2.
→
−
Example 103. Compute the flux of the vector field F = (x2 , y 2 , z)
through the sphere x2 +y 2 +z 2 = 1, from exterior to interior (see Fig.5).
Figure 5
We simply apply Gauss formula (95):
ZZ Z
→
−
Φ(S) ( F ) = − (2x + 2y + 1)dxdydz.
x2 +y 2 +z 2 ≤1
see that σρ ∂u
∂t
−K∆u is identical to zero on Ω (prove this in general like
in the case of a simple integral!). Finally we get the famous parabolic
PDE (partial differential equation) of the heat flow:
∂u
= c2 ∆u,
∂t
K
where c2 = σρ . If for a small interval of time the heat flow does not
depend on time, we get ∆u = 0, which is called the Laplace equation.
It is fundamental in many branches of Applied Mathematics. More
discussions on this subject can be find in any course of PDE.
3. Stokes Theorem
This fundamental result is a generalization of the Green theorem
(in plane), for the 3-dimensional case. This theorem says that in some
conditions a surface integral of the second type can be transformed into
a line integral and conversely.
A surface (S) ⊂ R3 has the curve Γ as its boundary or border
if Γ ⊂ (S), Γ is a piecewise smooth, connected and "closed" curve
(−
→r (a) = −→r (b), −
→
r : [a, b] → Γ ⊂ R3 ) and, for any point M0 of Γ, there
exists a small δ > 0 such that the intersection B(M0 ; r)∩(S), 0 < r ≤ δ,
between any ball (in R3 ) with centre at M0 and radius r, at most
δ, has the property that B(M0 ; r)Â [B(M0 ; r) ∩ (S)] is a connected
set. Moreover, if M0 ∈ (S) but M0 ∈ / Γ, then there exists at least
one ball B(M0 ; r), r > 0 such that B(M0 ; r)Â [B(M0 ; r) ∩ (S)] is the
disjoint union of two connected subsets of B(M0 ; r) (their intersections
is empty!).
A sphere S is not a surface with a border. But a semisphere is a
surface with a border Γ, the unique "big" circle on the sphere. Other
examples of surfaces with a border one can see in Fig.6, Fig.7 or Fig.8.
We must be careful! For instance, the cylinder x2 + y 2 = R2 , 0 ≤ z ≤ h
has not a (unique) border in our acceptance because its candidate for a
"border" is the nonconnected union between the following two circles:
x2 + y 2 = R2 , z = 0 and x2 + y 2 = R2 , z = h,where h 6= 0. If we add to
this last cylinder the "above" disc x2 + y 2 ≤ R2 , z = h, the obtained
surface has a unique (connected) border: x2 + y 2 = R2 , z = 0, thus it
is a surface with a border in our previous acceptance.
→
− →
− →
−
Let −→g : D → R3 , − →
g (u, v) = x(u, v) i + y(u, v) j + z(u, v) k , be a
piecewise smooth deformation and let (S) = − →g (D) be its corresponding
→
−
(direct oriented) surface with a border like above.. Let F (x, y, z) =
→
− →
− →
−
P (x, y, z) i +Q(x, y, z) j +R(x, y, z) k be a field of class C 1 defined on
a space domain which contains the surface (S). Let ∂D = γ be a direct
350 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
oriented closed piecewise smooth curve and let Γ = ∂(S) be the border
of (S) which is the image of γ through the deformation − →g . Notice that
Γ is "closed" and the orientation of Γ, which is that one inherited from
the direct orientation of γ in the uov-plane, is compatible with the
direct orientation of (S). This means that if we walk on the border Γ
in the direction indicated by its direct orientation, the direct face of
the surface (S) is always "on the left" (see Fig.6). The surface (S) of
Fig.6 is not simple w.r.t. Oz-axis (i.e. it has no parametrization of
the form z = z(x, y), (x, y) ∈ prxOy (S)), but we can divide it into two
nonoverlaping oriented surfaces (S1 ) and (S2 ) with the corresponding
borders ∂(S1 ) = [AEBCA] and ∂(S2 ) = [F ACBF ] respectively (see
Fig.6). It will be easy to prove that if the bellow Stokes formula is
true for simple surfaces (S1 ) and (S2 ), then it is true for the entire
(S) = (S1 ) ∪ (S2 ) (see this proof bellow!). This is why we shall prove
this basic formula only for a simple surface (S) w.r.t. Oz-axis.
Figure 6
→
−
− Let us recall that curl F is a new vector field associated to the field
→
F as follows:
⎛ −→ − → − → ⎞
→
− i j k
(3.1) curl F = det ⎝ ∂ ∂ ∂ ⎠=
∂x ∂y ∂z
P Q R
3. STOKES THEOREM 351
µ ¶ µ ¶ µ ¶
∂R ∂Q − → ∂P ∂R − → ∂Q ∂P − →
= − i + − j + − k.
∂y ∂z ∂z ∂x ∂x ∂y
Recall that the normal versor can be computed by the formula
µ ¶
→
− A B C
n = √ ,√ ,√
A2 + B 2 + C 2 A2 + B 2 + C 2 A2 + B 2 + C 2
√
and dσ = A2 + B 2 + C 2 dudv. Let us put
A B C
nx = √ , ny = √ , nz = √ .
A2 + B 2 + C 2 A2 + B 2 + C 2 A2 + B 2 + C 2
Theorem 96. (Stokes formula) Let D, (S), − →g , Γ = ∂(S), etc. be
the above notation and hypotheses. Then
I I I
→ −
− →
(3.2) (curl F ) · n dσ = P dx + Qdy + Rdz.
(S) Γ
→
−
The first surface integral is the flux of the curl F through the surface
→
−
(S). The last line integral is the circulation (or the work) of F on the
→
−
oriented border Γ of (S). Stokes formula says that this flux of curl F
→
−
and the circulation of F are equal.
→
−
Proof. Let us put the expression of curl F from (3.1) on the left
side of formula (3.2):
I I I I µ ¶ µ ¶
→ −
− → ∂R ∂Q ∂P ∂R
(curl F ) · n dσ = [ − nx + − ny +
∂y ∂z ∂z ∂x
(S) (S)
µ ¶ I I µ ¶
∂Q ∂P ∂P ∂P
+ − nz ]dσ = ny − nz dσ+
∂x ∂y ∂z ∂y
(S)
I I µ ¶ I I µ ¶
∂Q ∂Q ∂R ∂R
+ nz − nx dσ + nx − ny dσ.
∂x ∂z ∂y ∂x
(S) (S)
I
To prove that this last sum is equal to P dx + Qdy + Rdz it is enough
Γ
to prove the following three equalities:
I I ³ ´ I
∂P ∂P
a) n − ∂y nz dσ = P dx,
∂z y
(S) Γ
352 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
I I I
¡ ∂Q ∂Q
¢
b) ∂x
nz − n
∂z x
dσ = Qdy,
(S) Γ
I I ³ ´ I
∂R ∂R
c) n
∂y x
− n
∂x y
dσ = Rdz.
(S) Γ
Since the proof of each of these equalities are similar, we shall prove
only a). Now we can suppose that (S) is simple w.r.t. Oz-axis (other-
wise we decompose it into a finite simple surfaces like in Fig.6). Hence,
one can find for (S) an explicit parametrization w.r.t. z :
⎧
⎨ x=x
(S) : y=y , (x, y) ∈ Ωxy = prxOy (S)
⎩ z = z(x, y)
ZZ ∙ ¸
∂P ∂z ∂P
= − (x, y, z(x, y)) · − (x, y, z(x, y)) dxdy =
∂z ∂y ∂y
Ωxy
ZZ
∂
= − [P (x, y, z(x, y))] dxdy =
∂y
Ωxy
I I
Green
= P (x, y, z(x, y))dx = P (x, y, z)dx,
γ Γ
i.e. the right side of a). Here we just applied Green formula for the
domain Ωxy with its boundary γ (see Fig.7).
If (S) is not simple with respect to Oz-axis we reduce everything to
simple components of (S). ‘Let us apply this last idea for the surface
3. STOKES THEOREM 353
Figure 7
of Fig.6.
⎡ ⎤ ⎡ ⎤
I I I I I I
→ →
− ⎢ → → ⎥ ⎢
− → → ⎥
−
(curl F )·−
n dσ = ⎣ (curl F ) · −
n dσ ⎦+⎣ (curl F ) · −
n dσ ⎦ =
(S) (S1 ) (S2 )
⎡ ⎤
I I
⎢ ⎥
⎣ P dx + Qdy + Rdz + P dx + Qdy + Rdz ⎦ +
[AEB] [BCA]
⎡ ⎤
I I
⎢ ⎥
+⎣ P dx + Qdy + Rdz + P dx + Qdy + Rdz ⎦ .
[BF A] [ACB]
But I I
P dx + Qdy + Rdz = − P dx + Qdy + Rdz,
[BCA] [ACB]
so that
I I I
− →
→
(curl F ) · −
n dσ = P dx + Qdy + Rdz+
(S) [AEB]
I I
+ P dx + Qdy + Rdz = P dx + Qdy + Rdz.
[BF A] Γ
354 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
¤
I
Example 104. Let us compute the line integral I = x2 ydx +
Γ
y 2 dy + z 2 dz, where Γ is the oriented curve [O0 A0 B 0 C 0 O0 ] from Fig.8.
For this, let us consider the oriented surface
⎧
⎨ x=x
(S) : y = y , (x, y) ∈ [0, 1] × [0, 1],
⎩ z=1
I
+ x2 ydx + y 2 dy + z 2 dz+
[A0 B 0 ]:x=1,y=y,z=1,y∈[0,1]
I
+ x2 ydx + y 2 dy + z 2 dz+
[B 0 C 0 ]:x=x,y=1,z=1,x∈[0,1]−
3. STOKES THEOREM 355
Figure 8
I
+ x2 ydx + y 2 dy + z 2 dz =
[C 0 O0 ]:x=0,y=y,z=1,y∈[0,1]−
Z1 Z1 Z1
2 2 1
=0+ y dy − x dx − y 2 dy = − .
3
0 0 0
We just obtained the same result like that one obtained by using Stokes
formula. Hence, these last computation can also be viewed as a simple
verification.
→
−
We know that if F − : D → R3 if a conservative field of class C 1 on
→ →
−
a domain D, then curl F = 0 (see theorem 68) Conversely, if D is
a simple connected domain (any totally closed surface S ⊂ D can be
→
−
continuously shrunk to a point in D), then the field F of class C 1 is
→
−
conservative. We easily see that in order to prove that F is conserva-
I
→ →
−
tive, it is enough to prove that the line integral F · d− r = 0 on any
γ
→
−
"closed" piecewise smooth curve γ ⊂ D (construct a primitive for F of
I
(x,y,z)
→ −
−
the type: U(x, y, z) = F · d→
r ). Now, for any "closed" piecewise
(a,b,c)
smooth curve γ ⊂ D, it is intuitively clear (at least for the usual balls,
cylindrical domains, conical domains, paraboloidal domains, etc.) that
there exists a bounded, connected piecewise smooth surface T ⊂ D
356 9. GAUSS AND STOKES FORMULAS. APPLICATIONS
with the unique border γ (try to explain this existence at least for a
starred domain D, i.e. if there exists a point P0 ∈ D such that if P is
another point in D, then the whole segment [P0 P ] is included in D).
Then we can apply Stokes formula for the surface T with its border γ :
I I I
→ −
− →
(curl F ) · n dσ = P dx + Qdy + Rdz.
T γ
I
−
→ →
− →
−
But curl F = 0 , thus P dx + Qdy + Rdz = 0. Hence F is conserv-
γ
ative.
Sometimes Stokes formula cannot be easily applied, but this last
result can work easily.
I
Example 105. Let us compute the line integral I = x2 dx+y 2 dy+
Γ
z 2 dz, where Γ this time is the oriented curve [ABCD] of Fig.9.
Figure 9
→
− →
− →
− →
−
Since F = (x2 , y 2 , z 2 ), curl F = 0 and this implies that F is
I
conservative. Hence, the integral x2 dx + y 2 dy + z 2 dz does not depend
Γ
on the integration path which connects the points A and D. The simplest
path which connect A and D is the oriented segment [AD] : x = 0,
4. PROBLEMS AND EXERCISES 357
y = y, z = 0. Thus,
I ZR
2
x dx + y dy + z dz = y 2 dy = R3 .
2 2 2
3
Γ −R
→
− → −
− →
9. Use Stokes formula to compute the circulation of F = x j + y k
along the "closed" curve S, the intersection between the sphere x2 +
y 2 + z 2 = R2 and the plane x + y + z = 0, with the orientation given
by the direction of the vector (1, 1, 1).
→
− →
− →
− →
− →
−
10. Let F = yz i + 2xz j − x2 k . Compute the flux of F through
the ellipsoid 4x2 + y 2 + 4z 2 = 8, the exterior side and the circulation of
the same field along the intersection curve γ, the intersection between
this last ellipsoid and the plane z = 1.The projection of the orientation
of this curve on the xOy-plane is "from y to x", i.e. the clockwise
orientation.
CHAPTER 10
Figure 1
Lagrange
X
n
(1.6) = f (g(cj )) [x0 (dj ) + iy 0 (ej )] (tj − tj−1 ),
j=1
with
X
n
f (g(cj )) [x0 (cj ) + iy 0 (cj )] (tj − tj−1 ),
j=1
ZT
which is a Riemann sum of f (g(t))g 0 (t)dt. Hence this last integral can
T0
X
n
be well approximated with Riemann sums of the form f (ξ j )(zj −
j=1
Z
zj−1 ). This means that f is integrable on the arc ab and f (z)dz =
ab
ZT
f (g(t))g 0 (t)dt.
T0
Z
1
Example 106. Let us compute In = (z−z0 )n
dz for any nat-
|z−z0 |=r
ural number n = 0, 1, ... . The equation |z − z0 | = r describes a circle
with centre at z0 and radius r. It has the following parametrization
z½ = z0 + r exp(iθ), where θ ∈ [0, 2π]. It is equivalent to the real form
x = x0 + r cos θ
, θ ∈ [0, 2π]. Applying formula (1.5) we get:
y = y0 + r sin θ
Z2π ¯2π
1 iθ −n+1 eiθ(1−n) ¯¯
In = rie dθ = r i· = 0,
rn einθ i(1 − n) ¯0
0
Z2π
if n 6= 1 and I1 = i dθ = 2πi. Here we used the equality: e2πmi =
0
cos 2πm + i sin 2πm = 1 for any integer number m. Hence
Z ½
1 0, if n 6= 1
(1.7) dz =
(z − z0 )n 2πi, n = 1.
|z−z0 |=r
Z
We shall now express the complex integral f (z)dz as a complex
ab
number, with a real and an imaginary part. For this, let ξ j = η j + iθj ,
364 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
X
n
£ ¤
= u(η j , θj ) + iv(η j , θj ) [(x(tj ) − x(tj−1 )) + i(y(tj ) − y(tj−1 ))] =
j=1
X
n
£ ¤
= u(η j , θj )(x(tj ) − x(tj−1 )) − v(η j , θj )(y(tj ) − y(tj−1 )) +
j=1
X
n
£ ¤
+i u(η j , θj )(y(tj ) − y(tj−1 )) + v(η j , θj )(x(tj ) − x(tj−1 )) .
j=1
Since the real part of this last sum is a Riemann sum for the line
I
integral udx − vdy and since the imaginary part is a Riemann sum
ab
I
for the line integral vdx + udy, we find that:
ab
Z I I
(1.8) f (z)dz = udx − vdy + i vdx + udy.
ab ab ab
Z
Let us compute (3xy + y 2 i) dz, where γ is the oriented segment
γ
[AB] with A(0, 0) and B(1, 2). Since u(x, y) = 3xy and v(x, y) = y 2
can be easily computed, we use formula (1.8). A parametrization for
[AB] is x = x, y = 2x, x ∈ [0, 1]. Hence,
Z Z1 Z1
2 16
f (z)dz = (6x − 8x )dx + i (4x2 + 12x2 )dx = − + i .
2 2
3 3
ab 0 0
Figure 2
The simple connected property of D is necessarily. Indeed, for
instance Z
1
dz = 2πi 6= 0
z − z0
|z−z0 |=r
(see formula (1.7)) because D = CÂ{z0 } is not simple connected.
A main consequence of this Cauchy fundamental theorem is the pos-
sibility to construct a primitive function F (z) for an analytic function
I
f (z) defined on a simple connected domain D. The formula f (z)dz =
γ
0 for any piecewise smooth "closed" oriented curve γ ⊂ D, implies
ZB
like usually that the complex integral f (z)dz does not depend on
A
366 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
Z
f (ζ)dζ
segm[w0 w]
= ,
w − w0
where segm[w0 w] is the segment of the straight line which connect w0
and w (see Fig.3).
Figure 3
1. COMPLEX FUNCTIONS INTEGRATION 367
F (w) − F (w0 )
lim = f (w0 ),
w→w0 w − w0
F (w)−F (w0 )
let us evaluate now the distance between w−w0
and f (z) :
¯ ¯
¯ Z Z ¯
¯ ¯
¯ ¯
¯ f (ζ)dζ − f (w0 )dζ ¯
¯ ¯ ¯ ¯
¯ F (w) − F (w0 ) ¯ ¯segm[w0 w] segm[w0 w] ¯
¯ − f (w )¯= ≤
¯ w − w0
0 ¯ |w − w0 |
Z
|f (ζ) − f (w0 )| |dζ|
segm[w0 w]
≤ ≤ sup |f (ζ) − f (w0 )| .
|w − w0 | ζ∈segm[w0 w]
Since f is continuous
¯ on D, this¯ last supremum goes to zero when w →
¯ F (w)−F (w0 ) ¯
w0 . Hence, ¯ w−w0 − f (w0 )¯ → 0, when w → w0 , i.e. lim F (w)−F (w0 )
w−w0
=
w→w0
f (w0 ), i.e. F is differentiable at w0 and F 0 (w0 ) = f (w0 ).
What happens when D is not simple connected, i.e. if it has some
"gaps" or "holes"? We shall describe in the following a very simple
model of a n-connected domain Ω. We start with a complex domain D∗
and let us consider another simple connected bounded domain D ⊂ D∗
such that D = D ∪ ∂D ⊂ D∗ . Let D1 , D2 , ..., Dn−1 be n − 1 nonover-
laping (the intersection Di ∩ Dj , i 6= j, is empty or at most a piecewise
smooth curve) domains contained in D. We assume that the boundary
∂D = Γ+ 0 is a directly oriented piecewise smooth curve and the bound-
aries ∂D1 = Γ− −
1 , ..., ∂Dn−1 = Γn−1 are included in D and they are piece-
wise smooth inverse oriented curves (see Fig.4). Then Ω = DÂ∪n−1 j=1 D j
is said to be a n-connected domain. For n = 1 we get Ω = D itself, i.e.
a simple connected domain. For n = 2, Ω = DÂD1 and we obtain a
double connected domain (with a "hole"), etc. To obtain from Ω a new
simple connected domain Ω∗ , it is enough to get out from it n−1 piece-
wise smooth curves γ 1 , ..., γ n−1 which connect ∂D with ∂D1 , ..., ∂Dn−1
respectively, traced twice in the opposite directions (see Fig.4).
Hence Ω∗ , which has as its boundary the union between
∂D, ∂D1 , ..., ∂Dn−1 and γ 1 , ..., γ n−1 traced twice each of them, is a
simple connected domain.
368 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
Figure 4
Theorem 98. (n-connected Cauchy fundamental theorem) With
these notation and hypotheses, let f : D∗ Â ∪n−1
j=1 Dj → C be a con-
tinuous function which is analytic on Ω. Then,
Z Z Z Z
(1.9) f (z)dz = f (z)dz + f (z)dz + ... + f (z)dz.
Γ+
0 Γ+
1 Γ+
2 Γ+
n−1
Figure 5
or
Z n−1 Z
X
f (z)dz = f (z)dz.
j=1
Γ+
0 Γ+
j
Z
def 1
The residue res(f, zj ) = 2πi
f (z)dz is well defined because if
|z−zj |=rj
we diminish the radius rj of the disc to another one 0 < rj∗ < rj
then, on the annulus {z ∈ C : rj∗ < |z − zj | < rj } the function f (z)
is analytic and, applying the first statement of this theorem, we get
Z Z
1 1
that 2πi f (z)dz = 2πi f (z)dz, i.e. the residue res(f, zj ) is
|z−zj |=rj |z−zj |=rj∗
well defined. The equality (1.10) is the particular case of the general
Cauchy formula (1.9), where we take the discs |z − zj | < rj for the
domains Dj . ¤
Remark 37. Let H be a complex domain, let D be a bounded do-
main included in H such that ∂D = Γ is a direct oriented, continuous
and piecewise smooth curve. Let f be an analytic function on H, except
a finite number of points z1 , z2 , ...zm inside D and a finite number of
points w1 , w2 , ..., ws which are on ∂D = Γ (see Fig.6). If in a small
neighborhood Vj of wj , which does not contain any other wt , the curve
Γ is smooth, then we say that the angle in radians attached to wj is
αj = π. If wh is an angular point on Γ and the angle in radians between
the two distinct tangent lines at wh to Γ is αh , we attach this αh to wh
370 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
Figure 6
X
∞
2 n
(1.15) +a2 (z − z0 ) + ... + an (z − z0 ) + ... = an (z − z0 )n .
n=−∞
points at which the function is not analytic) of f (z). We easily see that
z0 = 0 is an essential singular point (see the definition bellow) of f
because sin 1z = z1 − 3!z1 3 + ... and (z2 +2)
1
2 is analytic (differentiable) in
√ √
a small neighborhood of z0 = 0. The points z1 = i 2 and z2 = −i 2
are
√ poles of order 2 because, for instance, the function √ g(z) = (z −
i 2)2 f (z) is analytic in a small neighborhood √ of i 2 which does not
contain the points z0 = 0 and z2 = −i 2, etc. It is clear enough
1. COMPLEX FUNCTIONS INTEGRATION 373
and
1 1
res(h, 1) = lim = .
z→1 z 2 (z + 1) 2
Hence, J = π2 i · 0 + π4 i · 1
2
= πi
8
.
If an infinite number of negative indexed terms a−n are distinct of
zero, we say that z0 is an essential point. If all a−n are zero, i.e. if
(1.18) f (z) = a0 + a1 (z − z0 ) + ... + an (z − z0 )n + ...
we say that f is analytic at z0 and z0 is a regular point of f . In this last
situation we call the expansion on the right side, a Taylor expansion.
It can be proved that any analytic function f in a disc B(z0 , R) has a
Taylor expansion of the type (1.18). Let us evaluate the coefficients an
in this last case. For this, let us take a circle γ r with centre at z0 and
an arbitrary radius r > 0 and let us integrate terms by terms:
Z X∞ Z
f (z)
n+1
dz = ak (z − z0 )k−n−1 dz = an · 2πi,
(z − z0 ) k=0
γr γr
Z Z
k−n−1
because of formulas (1.7) ( (z − z0 ) dz = 0 for k < n and (z −
γr γr
k−n−1
z0 ) dz = 2πi for k = n) and because of the Cauchy fundamental
Z
k−n−1
theorem (97) for k > n (then (z − z0 ) is analytic and so (z −
γr
z0 )k−n−1 dz = 0). Hence,
Z
1 f (z)
an = dz.
2πi (z − z0 )n+1
γr
Since an = n!1 f (n) (z0 ) (compute this n-th derivate in the Taylor expan-
sion (1.18)), one obtains:
Z
(n) n! f (z)
(1.20) f (z0 ) = dz.
2πi (z − z0 )n+1
γ
Z Z
00 1·2 f (z) (n) n! f (z)
f (w) = 3 dz, ..., f (w) = dz,
2πi (z − w) 2πi (z − w)n+1
|z−z0 |=r |z−z0 |=r
i.e. we just proved formula (1.20) and the amazing fact that if f (z)
in differentiable "once" on a simple connected domain D, then it has
derivatives of an arbitrary order on D.
Now, let us come back to the Taylor expansion (1.18) of an analytic
function f (z) in a disc B(z0 , R) and let r < R. Let us evaluate the
Z
1 f (z)
coefficient an = 2πi (z−z0 )n+1
dz, n = 0, 1, 2, ... :
γ: |z−z0 |=r
Z
1 1 1 1 M
|an | ≤ M |dz| = M · n+1 · 2πr = n ,
2π rn+1 2π r r
γ: |z−z0 |=r
Z2π
cos mx
For instance, if we want to compute the coefficient am = 2+sin x
dx of
0
1
the Fourier expansion of the function f (x) = 2+sin x
, x ∈ [0, 2π] (see
"Fourier series" chapter in any course of Advanced Mathematics), the
best idea is to use the above change of variables: z = eix . Thus,
Z
1 1 1 + z 2m 2iz 1
am = · m
· 2 · dz =
i 2 z z + 4iz − 1 z
|z|=1
Z Z
1 zm
= dz + dz.
z m (z 2 + 4iz − 1) z 2 + 4iz − 1
|z|=1 |z|=1
Z∞
b) Integrals of the form f (x)dx
−∞
These improper integrals of the first type can be easily computed
by using residues formula. First of all we need an auxiliary result.
Theorem 101. Let f (z) be an analytic function everywhere in the
upper half-plane {z ∈ C : Im z > 0}, except a finite number of points
z1 , z2 , ..., zn . We assume that in a neighborhood |z| > R0 > 0 of ∞ the
function f (z) is bounded as follows: |f (z)| < |z|M1+δ , where M > 0 and
δ > 0. Then
Z
(2.3) lim f (z)dz = 0,
R→∞
0
CR
Figure 7
Now we can prove the basic result which will help us to compute
improper integrals just announced.
Z∞ X
n
(2.4) f (x)dx = 2πi res(f, zk ),
−∞ k=1
where z1 , z2 , ..., zn are all the singular points (at which f is not analyti-
P (z)
cal) of f in the upper half-plane Im z > 0. In particular, if f (z) = Q(z)
is a rational function which satisfies the supplementary conditions that
deg Q(z) ≥ deg P (z) + 2 and that the equation Q(z) = 0 has no real
380 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
root, then
Z∞ X
P (x)
(2.5) dx = 2πi res(f, zk )
Q(x)
−∞ Re zk >0 and Q(zk )=0
Proof. Let us take R0 large enough such that |zk | < R0 for any
k = 1, 2, ..., n. Now we consider in the upper half-plane the closed
contour [−R, R] ∪ CR0 (see Fig.7). We apply the residue formula and
find:
ZR Z X
n
f (x)dx + f (z)dz = 2πi res(f, zk ).
−R 0 k=1
CR
Since the conditions of theorem 101 are satisfied, taking limit when
R → ∞, we get exactly formula (2.4).
P (x)
The last statement is true because, f (z) = Q(x) together with the
above restrictions on the polynomials P (z), Q(z), satisfies the same
conditions of theorem 101. ¤
R∞
For instance, we know that the integral I = −∞ x¯41+1 dx¯ is conver-
gent but its computation is not so easy. We see that ¯ z41+1 ¯ ≤ |z|13 , for
|z| large enough, i.e. R0 = 1, for instance, M = 1 and δ = 2 > 0 in
theorem 101. Thus we can apply the last theorem. For this we need to
find all the zeros of z 4 + 1 = 0 in the half-plane Im z > 0. The solutions
π+2kπ
of the equation z 4 + 1 = 0 are z0,1,2,3 = ei 4 , k = 0, 1, 2, 3. But only
π 3π
z0 = ei 4 and z1 = ei 4 are in the upper half-plane. Thus,
∙ µ ¶ µ ¶¸
1 i π4 1 i 3π
I = 2πi res 4 ,e + res 4 ,e 4 .
z +1 z +1
But, there is a problem! Formula (1.17) is not practicable here! Why?
Let us see:
µ ¶ ∙ ¸
1 i π4 1 1
res 4 ,e = lim (z − z0 ) 4 = .
z +1 z→z0 z +1 (z0 − z1 )(z0 − z2 )(z0 − z3 )
To compute this last number is not so easy. But, let us look carefully
to the general formula 1.17 for the case of a simple pole z0 , i.e. for
P (z)
k = 1 and for f (z) = Q(z) , a rational function:
Zπ
(2.8) = Rμ(R) e−aR sin θ dθ,
0
because
¯ iaz ¯ ¯¯ iaR·eiθ ¯¯ ¯ iaR(cos θ+i sin θ) ¯ ¯ iaR cos θ ¯ ¯ −aR sin θ) ¯
¯e ¯ = ¯e ¯ = ¯e ¯ = ¯e ¯ ¯e ¯=
¯ ¯ ¯ ¯
= |cos(aR cos θ) + i sin(aR cos θ)| ¯e−aR sin θ) ¯ = ¯e−aR sin θ) ¯ .
382 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
and
π
Z2 ¯π
2aR π 2aR ¯ 2 π £ −aR ¤
I ≤ 2 e− π θ dθ = − e− π θ ¯ = − e −1 .
aR 0 aR
0
Since
¯ ¯
¯Z ¯ Zπ
¯ ¯ π£ ¤
¯ iaz ¯
¯ e f (z)dz ¯ ≤ Rμ(R) e−aR sin θ dθ ≤ μ(R) 1 − e−aR
¯ ¯ a
¯C 0 ¯ 0
R
when R → ∞. ¤
Remark 38. a) If a < 0 and f (z) satisfies all the conditions which
appear in the statement of the theorem 103 for the lower half-plane
Im z ≤ 0, then the formula 2.7 is again true for any semicircular arc
CR0 in the lower half of the xOy-plane. b) Similar assertions hold for
2. APPLICATIONS OF RESIDUES FORMULA 383
Figure 8
Z π
Zr
iz 2 z=ρei 4 i π4 2
e dz = −e e−ρ dρ.
[BO] 0
Hence
Z∞ Z∞ √
2 2 π
cos x dx = sin x dx = √ .
2 2
0 0
386 10. SOME REMARKS ON COMPLEX FUNCTIONS INTEGRATION
Exam samples
1. June, 2010
2. June, 2010
√ √
1. The graphic of y = x 4 x2 + 1, x ∈ [0, 1] rotates around Ox-
axis. Find the volume of the resulting
R∞ solid.
1
2. Prove that the integral 0 (x2 +1)(x2 +4) dx is convergent and then
compute it.
√ 3. Find 2the area of the plane surface bounded by the curves: y =
10x and y = 10x.
ZZ Z
4. Compute zdxdydz. Draw this domain and sup-
x2 +y 2 +z 2 ≤9, z≥0
ply with a mechanical interpretation the obtained result.
3. June, 2010
R∞ tan−1 x
1. Prove that I = 1 x2
is convergent
dx ½ and compute it.
x = t2
2. Compute the length of the curve Γ : , t ∈ [0, 1].
y = t3
3. Let O(0, 0), A(1, 1), B(1, −1) and let the lamina OAB with the
density function f (x) = x. Find its mass and the coordinates of its
centre of mass.
387
388 A. EXAM SAMPLES
4. Find the
½ coordinates of the centre of mass for the total surface
x2 + y 2 = 4z 2
of the cone: .
0≤z≤2
4. June, 2010
3
1. The curve y = xe−x , x ∈ [0, ∞) rotates around the Ox-axis.
Find the volume of the obtained solid. First of all verify if this volume
is finite.
2. Starting with the formula
Z 1
Γ(x)Γ(y)
tx−1 (1 − t)y−1 dt = , x > 0, y > 0,
0 Γ(x + y)
Rπ
compute J = 02 sin6 x cos6 xdx.
3. Let A(2, 1) and B(1, 0). The homogenous lamina OAB of density
2, rotates around Ox-axis. Find its moment of inertia w.r.t. Ox-axis.
ZZZ
2 2 2
4. Let D : x + y + z ≤ 9, z ≥ 0. Compute I = z 2 dxdydz.
D
5. September, 2010
√ √
1. Let A(2, 2), B(1, 0) and C(2, − 2). Let γ be the arc of the
parabola y 2 = x which connect A and C. Let [ABC] the polygonal line
which passes through A, B and C. Compute the area bounded by γ
and [ABC].
→
− →
−
2. Find the work of the field F (x, y) = xy j along the polygonal
arc [OABO], with this orientation, where A(1, 2) and B(3, 0).
3. Compute the mass of the lamina D : x2 + y 2 ≤ 9, x ≥ 0, y ≤ 0,
if the density function is f (x, y) = |y| .
4. Find the coordinates of the centre of mass for the solid bounded
by the sphere x2 + y 2 + z 2 = 9 and inside the cylinder x2 + y 2 = 4,
z = 0, if the density function is f (x, y, z) = z.
6. September, 2010
R∞
1. Prove that I = 0 √ 1 3 dx is convergent and then compute
( x+1)
it.
2. Draw the domain D : x2 + y 2 ≤ 4, x2 + (y − 1)2 ≥ 1 and then
find the centre of mass of D (the density function is considered to be
f = 1). ¡ ¢ R∞ 3
3. If a = Γ 23 , compute K = 0 x10 e−x dx as a function of a.
4. Find the volume of the domain D : x2 +y 2 ≤ 2z, x2 +y 2 +z 2 ≤ 3.
APPENDIX B
Basic antiderivatives
R
Prove all the following formulas (here f (x)dx means a primitive
of f (x)):
⎧
Z ⎨ 1
α+1
xα+1 ,
if α 6= −1.
α
(0.1) x dx = ln x, if α = −1 and x > 0 .
⎩
ln(−x), if α = −1 and x < 0
Z
ax
(0.2) ax dx = , if a > 0, a 6= 1.
ln a
Z ¯ ¯
1 1 ¯¯ x − a ¯¯
(0.3) = ln , a 6= 0, x 6= ±a.
x2 − a2 2a ¯ x + a ¯
Z
1 1 x
(0.4) 2 2
dx = tan−1 , a 6= 0.
x +a a a
Z ¯ √ ¯
1 ¯ ¯
(0.5) √ dx = ln ¯x + x2 + a2 ¯ , a 6= 0.
2
x +a 2
Z
1 x
(0.6) √ dx = sin−1 , a 6= 0, a2 − x2 > 0.
2
a −x 2 a
Z √
x√ 2 a2 x
(0.7) a2 − x2 dx = a − x2 + sin−1 , a 6= 0, a2 − x2 > 0.
2 2 a
Z √
x√ 2 a2 √
(0.8) x2 ± a2 dx = x ± a2 ± ln(x + x2 ± a2 ), a 6= 0.
2 2
Z
(0.9) sin xdx = − cos x.
Z
(0.10) cos xdx = sin x.
389
390 B. BASIC ANTIDERIVATIVES
Z
1 π
(0.11) 2
dx = tan x, x 6= (2k + 1) .
cos x 2
Z
1
(0.12) dx = − cot x, x 6= kπ.
sin2 x
Z
π
(0.13) tan xdx = − ln |cos x| , x 6= (2k + 1) .
2
Z
(0.14) cot xdx = ln |sin x| , x 6= kπ.
Z ¯
1 ¯ x ¯¯
(0.15) dx = ln ¯tan ¯ , x 6= kπ.
sin x 2
Z ¯ ³ π x ´¯
1 ¯ ¯ π
(0.16) dx = ln ¯tan − ¯ , x 6= (2k + 1) .
cos x 4 2 2
Z
(0.17) sinh xdx = cosh x.
Z
(0.18) cosh dx = sinh x.
Z
1 sinh x
(0.19) 2 dx = tanh x = .
cosh x cosh x
Z
1 cosh x
(0.20) 2 dx = − coth x = − .
sinh x sinh x
Index
397