Landen's transformation is a mapping of the parameters of an elliptic integral, useful for the efficient numerical evaluation of elliptic functions. It was originally due to John Landen and independently rediscovered by Carl Friedrich Gauss.[1]
The incomplete elliptic integral of the first kind is
F(\varphi\setminus\alpha)=F(\varphi,\sin\alpha)=
\varphi | |
\int | |
0 |
d\theta | |
\sqrt{1-(\sin\theta\sin\alpha)2 |
\alpha
\alpha0
\alpha1
\varphi0
\varphi1
(1+\sin\alpha1)(1+\cos\alpha0)=2
\tan(\varphi1-\varphi0)=\cos\alpha0\tan\varphi0
\begin{align} F(\varphi0\setminus\alpha0)&=(1+
-1 | |
\cos\alpha | |
0) |
F(\varphi1\setminus\alpha1)\\ &=\tfrac{1}{2}(1+\sin\alpha1)F(\varphi1\setminus\alpha1). \end{align}
k=\sin\alpha
k'=\cos\alpha
In Gauss's formulation, the value of the integral
I=
| ||||
\int | ||||
0 |
1 | |
\sqrt{a2\cos2(\theta)+b2\sin2(\theta) |
a
b
a1=
a+b | |
2 |
, b1=\sqrt{ab},
I1=\int
| ||||
0 |
1 | |||||||||||||||
|
I= | 1 | K\left( |
a |
\sqrt{a2-b2 | |
I | K\left( | ||||
|
a-b | |
a+b |
\right).
K\left( | \sqrt{a2-b2 |
I1=I
The transformation may be effected by integration by substitution. It is convenient to first cast the integral in an algebraic form by a substitution of
\theta=\arctan(x/b)
d\theta=(\cos2(\theta)/b)dx
I=\int
| ||||
0 |
1 | |
\sqrt{a2\cos2(\theta)+b2\sin2(\theta) |
A further substitution of
x=t+\sqrt{t2+ab}
\begin{align}I&=\int
infty | |
0 |
1 | |
\sqrt{(x2+a2)(x2+b2) |
This latter step is facilitated by writing the radical as
\sqrt{(x2+a2)(x2+b2)}=2x\sqrt{t2+\left(
a+b | |
2 |
\right)2}
and the infinitesimal as
dx=
x | |
\sqrt{t2+ab |
so that the factor of
x
If the transformation is iterated a number of times, then the parameters
a
b
a
b
\operatorname{AGM}(a,b)
I=\int
| ||||
0 |
1 | |
\sqrt{a2\cos2(\theta)+b2\sin2(\theta) |
The integral may also be recognized as a multiple of Legendre's complete elliptic integral of the first kind. Putting
b2=a2(1-k2)
I=
1 | |
a |
\int
| ||||
0 |
1 | |
\sqrt{1-k2\sin2(\theta) |
Hence, for any
a
K(k)=
\pi | |
2\operatorname{AGM |
(1,\sqrt{1-k2})}
By performing an inverse transformation (reverse arithmetic-geometric mean iteration), that is
a-1=a+\sqrt{a2-b2}
b-1=a-\sqrt{a2-b2}
\operatorname{AGM}(a,b)=\operatorname{AGM}\left(a+\sqrt{a2-b2},a-\sqrt{a2-b2}\right)
the relationship may be written as
K(k)=
\pi | |
2\operatorname{AGM |
(1+k,1-k)}
which may be solved for the AGM of a pair of arbitrary arguments;
\operatorname{AGM}(u,v)=
\pi(u+v) | |||||
|
.