The Jordan lemma (after Marie Ennemond Camille Jordan ) is an aid to function theory . It is used in conjunction with the residual theorem to compute integrals from real analysis.
statement
If and converges uniformly to zero for all in the upper half-plane , then applies
α
>
0
{\ displaystyle \ alpha> 0}
G
{\ displaystyle g}
|
z
|
→
∞
{\ displaystyle | z | \ to \ infty}
∫
K
R.
G
(
z
)
e
i
α
z
d
z
→
0
{\ displaystyle \ int _ {K_ {R}} g (z) \, e ^ {i \ alpha z} dz \ to 0}
for .
R.
→
∞
{\ displaystyle R \ to \ infty}
This also applies if is and also tends to zero evenly in the upper half-plane. The lemma for the lower half-plane can be formulated completely analogously.
α
=
0
{\ displaystyle \ alpha = 0}
z
⋅
G
(
z
)
{\ displaystyle z \ cdot g (z)}
application
Many improper integrals of the form , if they exist, can be calculated in the following way: One integrates on a closed semicircular curve , which arises when integrating first on the real axis from to and from there in the semicircular arc back to .
∫
-
∞
∞
f
(
z
)
d
z
{\ displaystyle \ textstyle \ int _ {- \ infty} ^ {\ infty} f (z) \, dz}
f
{\ displaystyle f}
γ
R.
{\ displaystyle \ gamma _ {R}}
-
R.
{\ displaystyle -R}
R.
{\ displaystyle R}
K
R.
{\ displaystyle K_ {R}}
-
R.
{\ displaystyle -R}
It is found that for the integral vanishes and thus
R.
→
∞
{\ displaystyle R \ to \ infty}
∫
K
R.
f
d
z
{\ displaystyle \ textstyle \ int _ {K_ {R}} f \, dz}
∮
γ
R.
f
d
z
=
∫
[
-
R.
,
R.
]
f
d
z
+
∫
K
R.
f
d
z
→
R.
→
∞
∫
R.
f
d
z
{\ displaystyle \ oint _ {\ gamma _ {R}} fdz = \ int _ {[- R, R]} f \, dz + \ int _ {K_ {R}} f \, dz {\ xrightarrow [{R \ to \ infty}] {\}} \ int _ {\ mathbb {R}} f \, dz}
applies.
After the residual theorem is then
∫
R.
f
d
z
=
lim
R.
→
∞
∮
γ
R.
f
d
z
=
2
π
i
∑
I.
m
z
>
0
R.
e
s
f
|
z
{\ displaystyle \ int _ {\ mathbb {R}} f \, dz = \ lim _ {R \ to \ infty} \ oint _ {\ gamma _ {R}} fdz = 2 \ pi i \ sum _ {\ mathrm {Im} z> 0} \ mathrm {Res} f | _ {z}}
.
In order to avoid recurring estimates for integrals of the form, one uses the Jordan lemma.
∫
K
R.
G
(
z
)
e
i
α
z
d
z
{\ displaystyle \ textstyle \ int _ {K_ {R}} g (z) \, e ^ {i \ alpha z} dz}
Examples
1st example
It be and . The Jordan Lemma is applicable here and it holds
G
(
z
)
=
1
1
+
z
2
{\ displaystyle g (z) = {\ tfrac {1} {1 + z ^ {2}}}}
f
(
z
)
=
G
(
z
)
e
i
α
z
{\ displaystyle f (z) = g (z) \, e ^ {i \ alpha z}}
lim
R.
→
∞
∫
K
R.
f
(
z
)
d
z
=
0.
{\ displaystyle \ lim _ {R \ to \ infty} \ int _ {K_ {R}} f (z) \, dz = 0.}
So it holds for the integral over the real axis
∫
R.
f
(
z
)
d
z
=
2
π
i
R.
e
s
f
|
i
=
π
e
-
α
{\ displaystyle \ int _ {\ mathbb {R}} f (z) \, dz = 2 \ pi i \, \ mathrm {Res} f | _ {i} = \ pi \, e ^ {- \ alpha} }
.
If you split up into real and imaginary parts with the help of Euler's identity , you get equality
e
i
α
z
{\ displaystyle e ^ {i \ alpha z}}
∫
-
∞
∞
cos
(
α
x
)
1
+
x
2
d
x
=
π
e
-
α
{\ displaystyle \ int _ {- \ infty} ^ {\ infty} {\ frac {\ cos (\ alpha x)} {1 + x ^ {2}}} \, dx = \ pi \, e ^ {- \ alpha}}
.
2nd example
Be it . Analogous to the 1st example is and thus
G
(
z
)
=
z
1
+
z
2
{\ displaystyle g (z) = {\ tfrac {z} {1 + z ^ {2}}}}
∫
R.
f
(
z
)
d
z
=
2
π
i
R.
e
s
f
|
i
=
i
π
e
-
α
{\ displaystyle \ textstyle \ int _ {\ mathbb {R}} f (z) \, dz = 2 \ pi i \, \ mathrm {Res} f | _ {i} = i \ pi \, e ^ {- \ alpha}}
∫
-
∞
∞
x
sin
(
α
x
)
1
+
x
2
d
x
=
π
e
-
α
{\ displaystyle \ int _ {- \ infty} ^ {\ infty} {\ frac {x \ sin (\ alpha x)} {1 + x ^ {2}}} \, dx = \ pi \, e ^ { - \ alpha}}
.
Proof of the Jordan Lemma
The integral can be written after substitution as . Estimation of the amount upwards results
I.
R.
: =
∫
K
R.
G
(
z
)
e
i
α
z
d
z
{\ displaystyle \ textstyle I_ {R}: = \ int _ {K_ {R}} g (z) \, e ^ {i \ alpha z} \, dz}
z
=
R.
e
i
φ
{\ displaystyle z = R \, e ^ {i \ varphi}}
∫
0
π
G
(
R.
e
i
φ
)
e
i
α
R.
e
i
φ
R.
e
i
φ
i
d
φ
{\ displaystyle \ textstyle \ int _ {0} ^ {\ pi} g \ left (Re ^ {i \ varphi} \ right) \, e ^ {i \ alpha Re ^ {i \ varphi}} \, R \ , e ^ {i \ varphi} \, i \, d \ varphi}
|
I.
R.
|
≤
R.
ε
R.
∫
0
π
e
-
α
R.
sin
φ
d
φ
{\ displaystyle | I_ {R} | \ leq R \, \ varepsilon _ {R} \ int _ {0} ^ {\ pi} e ^ {- \ alpha R \ sin \ varphi} \, d \ varphi}
with . It follows
ε
R.
: =
Max
z
∈
K
R.
|
G
(
z
)
|
{\ displaystyle \ textstyle \ varepsilon _ {R}: = \ max _ {z \ in K_ {R}} | g (z) |}
|
I.
R.
|
≤
2
R.
ε
R.
∫
0
π
2
e
-
α
R.
sin
φ
d
φ
{\ displaystyle | I_ {R} | \ leq 2R \, \ varepsilon _ {R} \ int _ {0} ^ {\ frac {\ pi} {2}} e ^ {- \ alpha R \ sin \ varphi} \, d \ varphi}
,
since the integrand is axially symmetric with respect to . According to Jordan’s inequality, is for all and therefore
e
-
α
R.
sin
φ
{\ displaystyle e ^ {- \ alpha R \ sin \ varphi}}
φ
=
π
2
{\ displaystyle \ varphi = {\ tfrac {\ pi} {2}}}
sin
(
φ
)
≥
2
π
φ
{\ displaystyle \ sin (\ varphi) \ geq {\ tfrac {2} {\ pi}} \, \ varphi}
φ
∈
[
0
,
π
2
]
{\ displaystyle \ varphi \ in \ left [0, {\ tfrac {\ pi} {2}} \ right]}
|
I.
R.
|
≤
2
R.
ε
R.
∫
0
π
2
e
-
α
R.
2
π
φ
d
φ
=
π
ε
R.
α
(
1
-
e
-
α
R.
)
≤
π
ε
R.
α
→
0
{\ displaystyle | I_ {R} | \ leq 2R \, \ varepsilon _ {R} \ int _ {0} ^ {\ frac {\ pi} {2}} e ^ {- \ alpha R {\ frac {2 } {\ pi}} \ varphi} \, d \ varphi = {\ frac {\ pi \, \ varepsilon _ {R}} {\ alpha}} \ left (1-e ^ {- \ alpha R} \ right ) \ leq {\ frac {\ pi \, \ varepsilon _ {R}} {\ alpha}} \ to 0}
for .
R.
→
∞
{\ displaystyle R \ to \ infty}
literature
<img src="https://de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">