improper_integrals
improper_integrals
HCM — 2023.
D EFINITION 1.1
Let f (x) be defined for every number x Ê a and be
Z b
integrable on every interval [a, b]. Then Φ(b) = f (x)d x
a
is defined on the interval [a, +∞). The limit
Z b
I = lim Φ(b) = lim f (x)d x (1)
b→+∞ b→+∞ a
D EFINITION 1.2
Z b
1 If the limit I = lim f (x)d x exists (as a finite
b→+∞ a
Z +∞
number) then the improper integrals f (x)d x are
a
called convergent.
D EFINITION 1.2
Z b
1 If the limit I = lim f (x)d x exists (as a finite
b→+∞ a
Z +∞
number) then the improper integrals f (x)d x are
a
called convergent.
Z b
2 If the limit I = lim f (x)d x does not exist or is
b→+∞ a
Z +∞
equal to ∞ then the improper integrals f (x)d x
a
are called divergent.
G EOMETRIC MEANING
Z +∞
If f (x) Ê 0, ∀x ∈ [a, +∞) and the integral f (x)d x is
Z a+∞
convergent then the improper integrals f (x)d x can
a
be interpreted as an area of the region
S = {(x, y)|x Ê a, 0 É y É f (x)}.
D EFINITION 1.3
Let f (x) be defined for every number x É b and be
Z b
integrable on every interval [a, b]. Then Ψ(a) = f (x)d x
a
is defined on the interval (−∞, b]. The limit
Z b
I = lim Ψ(a) = lim f (x)d x (2)
a→−∞ a→−∞ a
D EFINITION 1.4
Z b
1 If the limit I = lim f (x)d x exists (as a finite
a→−∞ a
Z b
number) then the improper integrals f (x)d x are
−∞
called convergent.
D EFINITION 1.4
Z b
1 If the limit I = lim f (x)d x exists (as a finite
a→−∞ a
Z b
number) then the improper integrals f (x)d x are
−∞
called convergent.
Z b
2 If the limit I = lim f (x)d x does not exist or is
a→−∞ a
Z b
equal to ∞ then the improper integrals f (x)d x are
−∞
called divergent.
G EOMETRIC MEANING
Z b
If f (x) Ê 0, ∀x ∈ (−∞, b], and the integral f (x)d x is
−∞
Z b
convergent then the improper integrals f (x)d x can
−∞
be interpreted as an area of the region
S = {(x, y)|x É b, 0 É y É f (x)}.
D EFINITION 1.5
If f (x) is defined on R and integrable on every interval
[a, b] then for every number c ∈ R, the improper integral of
type 1 of function f (x) on (−∞, +∞) is defined by
Z +∞ Z c Z +∞
f (x)d x = f (x)d x + f (x)d x. (3)
−∞ −∞ c
D EFINITION 1.5
If f (x) is defined on R and integrable on every interval
[a, b] then for every number c ∈ R, the improper integral of
type 1 of function f (x) on (−∞, +∞) is defined by
Z +∞ Z c Z +∞
f (x)d x = f (x)d x + f (x)d x. (3)
−∞ −∞ c
Z +∞
The improper integral of type 1 f (x)d x is convergent
Z c Z +∞ −∞
Z +∞ Z c Z +∞
f (x)d x = f (x)d x + f (x)d x =
−∞ −∞ c
³ ´ µ ¶
= F (c) − lim F (a) + lim F (b) − F (c)
a→−∞ b→+∞
Z +∞
The improper integral of type 1 f (x)d x is
−∞
convergent if and only if lim F (a) and lim F (b) exist as
a→−∞ b→+∞
finite numbers
Z +∞ ¯+∞
f (x)d x = F (+∞) − F (−∞) = F (x)¯ (6)
¯
−∞ −∞
E XAMPLE 1.1
Z +∞
Evaluate I = cos xd x.
0
E XAMPLE 1.1
Z +∞
Evaluate I = cos xd x.
0
SOLUTION
¯+∞
I = sin x ¯ = lim sin b − sin 0 = lim sin b.
¯
0 b→+∞ b→+∞
E XAMPLE 1.1
Z +∞
Evaluate I = cos xd x.
0
SOLUTION
¯+∞
I = sin x ¯ = lim sin b − sin 0 = lim sin b.
¯
0 b→+∞ b→+∞
E XAMPLE 1.2
−1
dx
Z
Evaluate I = 2
−∞ x
E XAMPLE 1.2
−1
dx
Z
Evaluate I = 2
−∞ x
SOLUTION
1 ¯¯−1
¯
1
I = − ¯ = 1 + lim = 1.
x −∞ a→−∞ a
E XAMPLE 1.3
+∞
dx
Z
Evaluate I =
−∞ 1 + x2
E XAMPLE 1.3
+∞
dx
Z
Evaluate I =
−∞ 1 + x2
SOLUTION
¯+∞
I = arctan x ¯ = lim arctan b − lim arctan a =
¯
−∞ b→+∞ a→−∞
π ³ π´
= − − = π.
2 2
So the given integral I is convergent.
E XAMPLE 1.4
Z +∞
2
Evaluate I = xe −x d x
0
E XAMPLE 1.4
Z +∞
2
Evaluate I = xe −x d x
0
SOLUTION
¯+∞
1 +∞ −x 2 1 1 1 1
Z
2 2
e d (−x 2 ) = − e −x ¯¯ = lim − e −b + =
¯
I =−
2 0 2 0 b→+∞ 2 2 2
So the given integral I is convergent.
E XAMPLE 1.5
For what values of α is the integral
Z +∞
dx
I=
1 xα
convergent?
E XAMPLE 1.5
For what values of α is the integral
Z +∞
dx
I=
1 xα
convergent?
If α 6= 1 then
µ ¶
1 1 1
I =− lim −
α−1 b→+∞ b α−1 1α−1
E XAMPLE 1.5
For what values of α is the integral
Z +∞
dx
I=
1 xα
convergent?
If α 6= 1 then
µ ¶
1 1 1
I =− lim −
α−1 b→+∞ b α−1 1α−1
1 1
If α > 1, then lim = 0. Therefore I = and so
b→+∞ b α−1 α−1
the integral I converges.
Dr.Huynh Thi Hong Diem (HCMUT-OISP) IMPROPER INTEGRALS HCM — 2023. 18 / 47
Type 1: Infinite intervals Newton-Leibniz’s Formula
1
If α < 1, then lim = +∞ and so the integral I
b→+∞ b α−1
diverges.
1
If α < 1, then lim = +∞ and so the integral I
b→+∞ b α−1
diverges.
If α = 1, then I = lim ln |b| − ln 1 = +∞ and so the
b→+∞
integral I diverges.
1
If α < 1, then lim = +∞ and so the integral I
b→+∞ b α−1
diverges.
If α = 1, then I = lim ln |b| − ln 1 = +∞ and so the
b→+∞
integral I diverges.
S UMMARY
+∞
dx
Z
1 If α > 1 then I = converges.
1 xα
1
If α < 1, then lim = +∞ and so the integral I
b→+∞ b α−1
diverges.
If α = 1, then I = lim ln |b| − ln 1 = +∞ and so the
b→+∞
integral I diverges.
S UMMARY
+∞
dx
Z
1 If α > 1 then I = converges.
1 xα
Z +∞
dx
2 If α É 1 then I = diverges.
1 xα
T HEOREM 1.1
Suppose that f and g are continuous functions on every
interval [a, b] ⊂ [a, +∞) with 0 É g (x) É f (x), ∀x Ê a.
Z +∞ Z +∞
1 If f (x)d x is convergent, then g (x)d x is
a a
convergent.
T HEOREM 1.1
Suppose that f and g are continuous functions on every
interval [a, b] ⊂ [a, +∞) with 0 É g (x) É f (x), ∀x Ê a.
Z +∞ Z +∞
1 If f (x)d x is convergent, then g (x)d x is
a a
convergent.
Z +∞ Z +∞
2 If g (x)d x is divergent then f (x)d x is
a a
divergent.
E XAMPLE 1.6
Determine whether the integral is convergent or divergent
1 + e −x
Z +∞
dx
1 x
E XAMPLE 1.6
Determine whether the integral is convergent or divergent
1 + e −x
Z +∞
dx
1 x
SOLUTION
1 + e −x 1
>
x x
+∞
1
Z
Since d x is divergent, so the integral
1 x
1 + e −x
Z +∞
d x is divergent.
1 x
T HEOREM 1.2
Suppose f (x), g (x) be integrable on every [a, b] ⊂ [a, +∞)
f (x)
and f (x), g (x) Ê 0, ∀x Ê a. Evaluate lim =λ
x→∞ g (x)
Z ∞ Z ∞
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
T HEOREM 1.2
Suppose f (x), g (x) be integrable on every [a, b] ⊂ [a, +∞)
f (x)
and f (x), g (x) Ê 0, ∀x Ê a. Evaluate lim =λ
x→∞ g (x)
Z ∞ Z ∞
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
Z ∞ Z ∞
2 If λ > 0 then g (x)d x and f (x)d x either
a a
converges or diverges.
T HEOREM 1.2
Suppose f (x), g (x) be integrable on every [a, b] ⊂ [a, +∞)
f (x)
and f (x), g (x) Ê 0, ∀x Ê a. Evaluate lim =λ
x→∞ g (x)
Z ∞ Z ∞
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
Z ∞ Z ∞
2 If λ > 0 then g (x)d x and f (x)d x either
a a
converges or diverges.
Z ∞ Z ∞
3 If λ = +∞ and g (x)d x diverges then f (x)d x
a a
diverges.
∞
I NDETERMINATE FORM
∞
xα
1 lim = 0 (a > 1)
x→+∞ a x
lnα x
2 lim = 0 (∀α > 0, β > 0)
x→+∞ x β
∞
I NDETERMINATE FORM
∞
xα
1 lim = 0 (a > 1)
x→+∞ a x
lnα x
2 lim = 0 (∀α > 0, β > 0)
x→+∞ x β
D EFINITION 1.6
Functions f (x) and g (x) are called equivalent as x → a, if
f (x)
lim =1 (7)
x→a g (x)
x→a
Denote by f (x) ∼ g (x).
E XAMPLE 1.7
Determine
Z +∞ whether the integral is convergent or divergent
dx
I= p
1 x 2 − 2x + 3
E XAMPLE 1.7
Determine
Z +∞ whether the integral is convergent or divergent
dx
I= p
1 x 2 − 2x + 3
1
We have p > 0, ∀x Ê 1 and
x 2 − 2x + 3
1 x→+∞ 1
p ∼ ,
x 2 − 2x + 3 x
+∞
dx
Z
where diverges, therefore I diverges.
1 x
E XAMPLE 1.8
Determine
Z +∞ whether the integral is convergent or divergent
(x + 1)d x
I= p
3
2 x 7 − 3x − 2
E XAMPLE 1.8
Determine
Z +∞ whether the integral is convergent or divergent
(x + 1)d x
I= p
3
2 x 7 − 3x − 2
x +1
We have p3
> 0, ∀x Ê 2 and
x 7 − 3x − 2
x +1 x→+∞ x 1
p ∼ = ,
3
x 7 − 3x − 2 x 7/3 x 4/3
+∞
dx
Z
where converges, therefore I converges.
2 x 4/3
Z b Z η
f (x)d x = lim− Φ(η) = lim− f (x)d x (8)
a η→b η→b a
D EFINITION 2.2
Z η
1 If lim− Φ(η) = lim− f (x)d x exists (as a finite
η→b η→b a
number) then the improper integral of type 2
Z b
f (x)d x converges.
a
D EFINITION 2.2
Z η
1 If lim− Φ(η) = lim− f (x)d x exists (as a finite
η→b η→b a
number) then the improper integral of type 2
Z b
f (x)d x converges.
a
Z η
2 If lim− Φ(η) = lim− f (x)d x does not exist or is equal
η→b η→b a
Z b
to ∞ then the improper integral of type 2 f (x)d x
a
diverges.
G EOMETRIC MEANING
Z b
If f (x) Ê 0, ∀x ∈ [a, b) and the integral f (x)d x is
aZ
b
convergent then the improper integrals f (x)d x can be
a
interpreted as an area of the region
S = {(x, y)|a É x < b, 0 É y É f (x)}, where x = b is the
vertical asymptote of the graph of function f (x)
Z b Z b
f (x)d x = lim+ Ψ(ξ) = lim+ f (x)d x (9)
a ξ→a ξ→a ξ
D EFINITION 2.4
Z b
1 If lim+ Ψ(ξ) = lim+ f (x)d x exists (as a finite
ξ→a ξ→a ξ
number) then the improper integral of type 2
Z b
f (x)d x converges.
a
D EFINITION 2.4
Z b
1 If lim+ Ψ(ξ) = lim+ f (x)d x exists (as a finite
ξ→a ξ→a ξ
number) then the improper integral of type 2
Z b
f (x)d x converges.
a
Z b
2 If lim+ Ψ(ξ) = lim+ f (x)d x does not exist or is equal
ξ→a ξ→a ξ
Z b
to ∞ then the improper integral of type 2 f (x)d x
a
diverges.
G EOMETRIC MEANING
Z b
If f (x) Ê 0, ∀x ∈ (a, b] and the integral f (x)d x is
aZ
b
convergent then the improper integrals f (x)d x can be
a
interpreted as an area of the region
S = {(x, y)|a < x É b, 0 É y É f (x)}, where x = a is the
vertical asymptote of the graph of function f (x)
Z b Z c Z b
f (x)d x = f (x)d x + f (x)d x (10)
a a c
Z b Z c Z b
f (x)d x = f (x)d x + f (x)d x (10)
a a c
D EFINITION 2.5
Z b
The improper integral of type 2 f (x)d x converges if
Z c Z ba
Z b ¯b −
f (x)d x = F (b − 0) − F (a) = F (x)¯ . (11)
¯
a a
Z b ¯b
f (x)d x = F (b) − F (a + 0) = F (x)¯ + . (12)
¯
a a
E XAMPLE 2.1
Z 1
dx
Evaluate I =
0 x
E XAMPLE 2.1
Z 1
dx
Evaluate I =
0 x
SOLUTION
1
We have lim = +∞. Therefore x = 0 is vertical
x→0+ x
asymptote.
E XAMPLE 2.1
Z 1
dx
Evaluate I =
0 x
SOLUTION
1
We have lim = +∞. Therefore x = 0 is vertical
x→0+ x
asymptote. Since
¯1
I = ln |x|¯ = ln 1 − lim ln |a| = +∞.
¯
0 a→0+
E XAMPLE 2.2
Z 1
arccos x
Evaluate I = p dx
−1 1 − x2
arccos x
SOLUTION We have lim p = +∞. Therefore
x→−1+ 1 − x2
x = −1 is vertical asymptote.
E XAMPLE 2.2
Z 1
arccos x
Evaluate I = p dx
−1 1 − x2
arccos x
SOLUTION We have lim p = +∞. Therefore
x→−1+ 1 − x2
x = −1 is vertical asymptote.In other hand, x = 1 is not
arccos x
vertical asymptote because lim p = 1.
x→1− 1 − x2
E XAMPLE 2.2
Z 1
arccos x
Evaluate I = p dx
−1 1 − x2
arccos x
SOLUTION We have lim p = +∞. Therefore
x→−1+ 1 − x2
x = −1 is vertical asymptote.In other hand, x = 1 is not
arccos x
vertical asymptote because lim p = 1. Since
x→1− 1 − x2
Z 1
1 2
¯1
I =− arccos xd (arccos x) = − · (arccos x)¯
¯
−1 2 −1
1 2 2 π2
= − (arccos 1 − lim arccos a) =
2 a→−1+ 2
so improper integral I is convergent.
Dr.Huynh Thi Hong Diem (HCMUT-OISP) IMPROPER INTEGRALS HCM — 2023. 40 / 47
Type 2: Infinity discontinuous integrands Newton-Leibniz’s formula
E XAMPLE 2.3
b
dx
Z
Evaluate I = , (a < b)
a (b − x)α
E XAMPLE 2.3
b
dx
Z
Evaluate I = , (a < b)
a (b − x)α
SOLUTION
η ¯η
dx 1
Z
−α+1 ¯
I = lim− =− lim (b − x) ¯ =
η→b a (b − x)α −α + 1 η→b − a
1 1
= lim− (b − η)−α+1 + (b − a)−α+1
α − 1 η→b −α + 1
E XAMPLE 2.3
b
dx
Z
Evaluate I = , (a < b)
a (b − x)α
SOLUTION
η ¯η
dx 1
Z
−α+1 ¯
I = lim− =− lim (b − x) ¯ =
η→b a (b − x)α −α + 1 η→b − a
1 1
= lim− (b − η)−α+1 + (b − a)−α+1
α − 1 η→b −α + 1
If α < 1 then lim− (b − η)−α+1 = 0.
η→b
E XAMPLE 2.3
b
dx
Z
Evaluate I = , (a < b)
a (b − x)α
SOLUTION
η ¯η
dx 1
Z
−α+1 ¯
I = lim− =− lim (b − x) ¯ =
η→b a (b − x)α −α + 1 η→b − a
1 1
= lim− (b − η)−α+1 + (b − a)−α+1
α − 1 η→b −α + 1
If α < 1 then lim− (b − η)−α+1 = 0.
η→b
If α > 1 then lim− (b − η)−α+1 = ∞.
η→b
If α = 1 then
Zη ¯η
dx
I = lim− = − lim− ln |b − x|¯ =
¯
η→b b−x η→b a
a
= − lim− ln |b − η| + ln(b − a) = ∞.
η→b
If α = 1 then
Zη ¯η
dx
I = lim− = − lim− ln |b − x|¯ =
¯
η→b b−x η→b a
a
= − lim− ln |b − η| + ln(b − a) = ∞.
η→b
S UMMARY
b
dx
Z
1 If α < 1, then improper integral converges.
a (b − x)α
If α = 1 then
Zη ¯η
dx
I = lim− = − lim− ln |b − x|¯ =
¯
η→b b−x η→b a
a
= − lim− ln |b − η| + ln(b − a) = ∞.
η→b
S UMMARY
b
dx
Z
1 If α < 1, then improper integral α
converges.
a (b − x)
Z b
dx
2 If α Ê 1, then improper integral α
diverges.
a (b − x)
T HEOREM 2.1
Suppose f (x), g (x) be integrable on every [a, η] ⊂ [a, b)
and not be bounded as x → b − . Moreover, for every
f (x)
x ∈ [a, b), we have 0 É f (x), 0 É g (x), lim− =λ
x→b g (x)
Z b Z b
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
T HEOREM 2.1
Suppose f (x), g (x) be integrable on every [a, η] ⊂ [a, b)
and not be bounded as x → b − . Moreover, for every
f (x)
x ∈ [a, b), we have 0 É f (x), 0 É g (x), lim− =λ
x→b g (x)
Z b Z b
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
Z b Z b
2 If λ > 0 then g (x)d x and f (x)d x either
a a
converges or diverges.
T HEOREM 2.1
Suppose f (x), g (x) be integrable on every [a, η] ⊂ [a, b)
and not be bounded as x → b − . Moreover, for every
f (x)
x ∈ [a, b), we have 0 É f (x), 0 É g (x), lim− =λ
x→b g (x)
Z b Z b
1 If λ = 0 and g (x)d x converges then f (x)d x
a a
converges.
Z b Z b
2 If λ > 0 then g (x)d x and f (x)d x either
a a
converges or diverges.
Z b Z b
3 If λ = +∞ and g (x)d x diverges then f (x)d x
a a
diverges.
Dr.Huynh Thi Hong Diem (HCMUT-OISP) IMPROPER INTEGRALS HCM — 2023. 43 / 47
Type 2: Infinity discontinuous integrands Some basic equivalent functions
n
x2
6 sinh x ∼ x, cosh x − 1 ∼
2
E XAMPLE 2.4
Determine whether the integral is convergent or divergent
cos2 x
Z 1
I= p
3
dx
0 1 − x2
E XAMPLE 2.4
Determine whether the integral is convergent or divergent
cos2 x
Z 1
I= p
3
dx
0 1 − x2
cos2 x cos2 x 1
lim− p
3
= lim− p
3
. 1/3
=∞
x→1 1 − x 2 x→1 1 + x (1 − x)
E XAMPLE 2.4
Determine whether the integral is convergent or divergent
cos2 x
Z 1
I= p
3
dx
0 1 − x2
cos2 x cos2 x 1
lim− p
3
= lim− p
3
. 1/3
=∞
x→1 1 − x 2 x→1 1 + x (1 − x)
cos2 x cos2 x 1 2
x→1− cos 1 1
p = p . ∼ p . .
1 + x (1 − x)1/3 2 (1 − x)1/3
3 3 3
1 − x2
1
We have α = < 1, therefore I converges.
3
E XAMPLE 2.5
Determine whetherp the integral is convergent or divergent
Z 1 3
ln(1 + x)
I= dx
0 e sin x − 1
E XAMPLE 2.5
Determine whetherp the integral is convergent or divergent
Z 1 3
ln(1 + x)
I= dx
0 e sin x − 1
p
ln(1 + 3 x) x 1/3 1
lim+ sin x = lim+ = lim+ 2/3 = ∞
x→0 e −1 x→0 x x→0 x
E XAMPLE 2.5
Determine whetherp the integral is convergent or divergent
Z 1 3
ln(1 + x)
I= dx
0 e sin x − 1
p
ln(1 + 3 x) x 1/3 1
lim+ sin x = lim+ = lim+ 2/3 = ∞
x→0 e −1 x→0 x x→0 x
p
ln(1 + 3 x) x→0+ x 1/3 1
∼ =
e sin x − 1 x x 2/3
2
Due to α = < 1 so I converges.
3