Elliptic Integration
Elliptic integrals are of the form ∫abaxbx yxⅆx, where yx is a polynomial of degree 3 or 4, and ax and bx are polynomials.
restart
Complete Elliptic Integrals
Maple recognizes the complete elliptic integrals
∫0π211−k2sint2 ⅆt (First Kind)
∫0π21−k2sint2 ⅆt (Second Kind)
∫0π211−a sint2 1−k2sint2 ⅆt (Third Kind)
where 0<k<1. We see this as follows:
First, we tell Maple that the usual parameter k lies between 0 and 1.
assume⁡0<k,k<1
Now we evaluate the above integrals:
∫0π211−k2⁢sin⁡t2ⅆt
∫0π21−k2⁢sin⁡t2ⅆt
∫0π211−a⁢sin⁡t2⁢1−k2⁢sin⁡t2ⅆt
EllipticK⁡k~
EllipticE⁡k~
EllipticPi⁡a,k~
The tilde (~) following each k reminds us of the assumption of restricted range.
Maple can integrate the same type of integrals by using a hyperbolic trig argument. For instance, the integral
∫arcsinh12arcsinh111+k2sinht2 ⅆt evaluates to:
∫arcsinh⁡12arcsinh⁡111+k2⁢sinh⁡t2ⅆt
EllipticF⁡12⁢2,−k~2+1−EllipticF⁡15⁢5,−k~2+1
As another example, to compute the value of the integral ∫arcsinh12arcsinh12+k2sinht2 ⅆt, we evaluate:
∫arcsinh⁡12arcsinh⁡12+k2⁢sinh⁡t2ⅆt
2⁢EllipticF⁡12⁢2,12⁢4−2⁢k~2−2⁢EllipticF⁡15⁢5,12⁢4−2⁢k~2−12⁢k~2⁢2⁢EllipticF⁡12⁢2,12⁢4−2⁢k~2+12⁢k~2⁢2⁢EllipticF⁡15⁢5,12⁢4−2⁢k~2+12⁢k~2⁢2⁢EllipticPi⁡12⁢2,1,12⁢4−2⁢k~2−12⁢k~2⁢2⁢EllipticPi⁡15⁢5,1,12⁢4−2⁢k~2
Elliptic Integration and assume
Because elliptic integration works for variable arguments that take advantage of the Maple assume facility, you can investigate formulae that cross branches of the square root function. The assumptions are the following:
assume⁡0<k,k<1,1<z,z<1k
Consider the following integral, ∫0z11−x21−k2 x2 ⅆx. Evaluate and assign to answer1:
answer1:=∫0z11−x2⁢1−k2⁢x2ⅆx
answer1:=−I⁢EllipticF⁡z~2−11−k~2⁢z~,1−k~2+EllipticK⁡k~
The next integral, ∫0z1−x21−k2 x2 ⅆx, assigned to answer2, evaluates to
answer2:=∫0z1−x2⁢1−k2⁢x2ⅆx
answer2:=13⁢z~⁢1−k~2⁢z~2−z~2+k~2⁢z~4−23⁢I⁢EllipticF⁡z~2−11−k~2⁢z~,1−k~2−I⁢−13⁢k~2−13⁢EllipticPi⁡z~2−11−k~2⁢z~,1−k~2,1−k~2+23⁢EllipticK⁡k~+−13⁢k~2−13⁢EllipticK⁡k~k~2−−13⁢k~2−13⁢EllipticE⁡k~k~2
We can do a simple check for the integral directly above by trying some values. For example, try k=12 and z=2.
check:=∫021−x2⁢1−1⁢x24ⅆx
check2:=subs⁡k=12,z=2,answer2
check:=−23⁢I⁢EllipticK⁡12⁢3+53⁢I⁢EllipticE⁡12⁢3−EllipticK⁡12+53⁢EllipticE⁡12
check2:=−23⁢I⁢EllipticF⁡12⁢4,14⁢3⁢4+512⁢I⁢EllipticPi⁡12⁢4,34,14⁢3⁢4−EllipticK⁡12+53⁢EllipticE⁡12
Compute check - check2. This should evaluate to zero.
simplify⁡check−check2
0
EllipticF, EllipticE, and EllipticPi
Integrals can be reduced to normal form in terms of the three Legendre elliptic functions: EllipticF, EllipticE, and EllipticPi. We begin by declaring some assumptions.
assume⁡0<k,k<1,k<z,z<1
Firstly, the EllipticF function is given by ∫0z11−x2 1−k2 x2 ⅆx.
∫0z11−x2⁢1−k2⁢x2ⅆx
EllipticF⁡z~,k~
Secondly, the EllipticE function has the form ∫0z1−k2x21−x2 ⅆx.
∫0z1−k2⁢x21−x2ⅆx
EllipticE⁡z~,k~
Thirdly, the EllipticPi function is of the form ∫0z11−a x2 1−k2 x2 1−x2 ⅆx.
∫0z11−a⁢x2⁢1−k2⁢x2⁢1−x2ⅆx
EllipticPi⁡z~,a,k~
For all of the above functions, the variable k must lie between 0 and 1. The Maple integrator facility reduces
∫0111+x4ⅆx
12⁢EllipticK⁡12⁢2
to a normal form expression. This can then be evaluated numerically to 30 (or more) digits.
evalf⁡,30
0.927037338650685959216925173600
Compare this to the Maple numerical integrator, and we see that the answers are the same (at least up to round errors).
evalf⁡
0.9270373387
In Maple, the integrator also recognizes the trigonometric form of these integrals:
∫abRsinθ,cosθ ysinθ, cosθ ⅆθ, where R is a rational function of sin and cos, and y is a quadratic polynomial in sin and cos.
See Also
assume, int
Return to Index for Example Worksheets
Download Help Document