ecterrab

9729 Reputation

20 Badges

15 years, 280 days

MaplePrimes Activity


These are Posts that have been published by ecterrab

Wirtinger Derivatives in Maple 2021

Generally speaking, there are two contexts for differentiating complex functions with respect to complex variables. In the first context, called the classical complex analysis, the derivatives of the complex components ( abs , argument , conjugate , Im , Re , signum ) with respect to complex variables do not exist (do not satisfy the Cauchy-Riemann conditions), with the exception of when they are holomorphic functions. All computer algebra systems implement the complex components in this context, and computationally represent all of abs(z), argument(z), conjugate(z), Im(z), Re(z), signum(z) as functions of z . Then, viewed as functions of z, none of them are analytic, so differentiability becomes an issue.

 

In the second context, first introduced by Poincare (also called Wirtinger calculus), in brief z and its conjugate conjugate(z) are taken as independent variables, and all the six derivatives of the complex components become computable, also with respect to conjugate(z). Technically speaking, Wirtinger calculus permits extending complex differentiation to non-holomorphic functions provided that they are ℝ-differentiable (i.e. differentiable functions of real and imaginary parts, taking f(z) = f(x, y) as a mapping "`ℝ`^(2)->`ℝ`^()").

 

In simpler terms, this subject is relevant because, in mathematical-physics formulations using paper and pencil, we frequently use Wirtinger calculus automatically. We take z and its conjugate conjugate(z) as independent variables, with that d*conjugate(z)*(1/(d*z)) = 0, d*z*(1/(d*conjugate(z))) = 0, and we compute with the operators "(∂)/(∂ z)", "(∂)/(∂ (z))" as partial differential operators that behave as ordinary derivatives. With that, all of abs(z), argument(z), conjugate(z), Im(z), Re(z), signum(z), become differentiable, since they are all expressible as functions of z and conjugate(z).

 

 

Wirtinger derivatives were implemented in Maple 18 , years ago, in the context of the Physics package. There is a setting, Physics:-Setup(wirtingerderivatives), that when set to true - an that is the default value when Physics is loaded - redefines the differentiation rules turning on Wirtinger calculus. The implementation, however, was incomplete, and the subject escaped through the cracks till recently mentioned in this Mapleprimes post.

 

Long intro. This post is to present the completion of Wirtinger calculus in Maple, distributed for everybody using Maple 2021 within the Maplesoft Physics Updates v.929 or newer. Load Physics and set the imaginary unit to be represented by I

 

with(Physics); interface(imaginaryunit = I)

 

The complex components are represented by the computer algebra functions

(FunctionAdvisor(complex_components))(z)

[Im(z), Re(z), abs(z), argument(z), conjugate(z), signum(z)]

(1)

They can all be expressed in terms of z and conjugate(z)

map(proc (u) options operator, arrow; u = convert(u, conjugate) end proc, [Im(z), Re(z), abs(z), argument(z), conjugate(z), signum(z)])

[Im(z) = ((1/2)*I)*(-z+conjugate(z)), Re(z) = (1/2)*z+(1/2)*conjugate(z), abs(z) = (z*conjugate(z))^(1/2), argument(z) = -I*ln(z/(z*conjugate(z))^(1/2)), conjugate(z) = conjugate(z), signum(z) = z/(z*conjugate(z))^(1/2)]

(2)

The main differentiation rules in the context of Wirtinger derivatives, that is, taking z and conjugate(z) as independent variables, are

map(%diff = diff, [Im(z), Re(z), abs(z), argument(z), conjugate(z), signum(z)], z)

[%diff(Im(z), z) = -(1/2)*I, %diff(Re(z), z) = 1/2, %diff(abs(z), z) = (1/2)*conjugate(z)/abs(z), %diff(argument(z), z) = -((1/2)*I)/z, %diff(conjugate(z), z) = 0, %diff(signum(z), z) = (1/2)/abs(z)]

(3)

Since in this context conjugate(z) is taken as - say - a mathematically-atomic variable (the computational representation is still the function conjugate(z)) we can differentiate all the complex components also with respect to  conjugate(z)

map(%diff = diff, [Im(z), Re(z), abs(z), argument(z), conjugate(z), signum(z)], conjugate(z))

[%diff(Im(z), conjugate(z)) = (1/2)*I, %diff(Re(z), conjugate(z)) = 1/2, %diff(abs(z), conjugate(z)) = (1/2)*z/abs(z), %diff(argument(z), conjugate(z)) = ((1/2)*I)*z/abs(z)^2, %diff(conjugate(z), conjugate(z)) = 1, %diff(signum(z), conjugate(z)) = -(1/2)*z^2/abs(z)^3]

(4)

For example, consider the following algebraic expression, starting with conjugate

eq__1 := conjugate(z)+z*conjugate(z)^2

conjugate(z)+z*conjugate(z)^2

(5)

Differentiating this expression with respect to z and conjugate(z) taking them as independent variables, is new, and in this example trivial

(%diff = diff)(eq__1, z)

%diff(conjugate(z)+z*conjugate(z)^2, z) = conjugate(z)^2

(6)

(%diff = diff)(eq__1, conjugate(z))

%diff(conjugate(z)+z*conjugate(z)^2, conjugate(z)) = 1+2*z*conjugate(z)

(7)

Switch to something less trivial, replace conjugate by the real part ReNULL

eq__2 := eval(eq__1, conjugate = Re)

Re(z)+z*Re(z)^2

(8)

To verify results further below, also express eq__2 in terms of conjugate

eq__22 := simplify(convert(eq__2, conjugate), size)

(1/4)*(z^2+z*conjugate(z)+2)*(z+conjugate(z))

(9)

New: differentiate eq__2 with respect to z and  conjugate(z)

(%diff = diff)(eq__2, z)

%diff(Re(z)+z*Re(z)^2, z) = 1/2+Re(z)^2+z*Re(z)

(10)

(%diff = diff)(eq__2, conjugate(z))

%diff(Re(z)+z*Re(z)^2, conjugate(z)) = 1/2+z*Re(z)

(11)

Note these results (10) and (11) are expressed in terms of Re(z), not conjugate(z). Let's compare with the derivative of eq__22 where everything is expressed in terms of z and conjugate(z). Take for instance the derivative with respect to z

(%diff = diff)(eq__22, z)

%diff((1/4)*(z^2+z*conjugate(z)+2)*(z+conjugate(z)), z) = (1/4)*(2*z+conjugate(z))*(z+conjugate(z))+(1/4)*z^2+(1/4)*z*conjugate(z)+1/2

(12)

To verify this result is mathematically equal to (10) expressed in terms of Re(z) take the difference of the right-hand sides

rhs((%diff(Re(z)+z*Re(z)^2, z) = 1/2+Re(z)^2+z*Re(z))-(%diff((1/4)*(z^2+z*conjugate(z)+2)*(z+conjugate(z)), z) = (1/4)*(2*z+conjugate(z))*(z+conjugate(z))+(1/4)*z^2+(1/4)*z*conjugate(z)+1/2)) = 0

Re(z)^2+z*Re(z)-(1/4)*(2*z+conjugate(z))*(z+conjugate(z))-(1/4)*z^2-(1/4)*z*conjugate(z) = 0

(13)

One quick way to verify the value of expressions like this one is to replace z = a+I*b and simplify "assuming" a andNULLb are realNULL

`assuming`([eval(Re(z)^2+z*Re(z)-(1/4)*(2*z+conjugate(z))*(z+conjugate(z))-(1/4)*z^2-(1/4)*z*conjugate(z) = 0, z = a+I*b)], [a::real, b::real])

a^2+(a+I*b)*a-(1/2)*(3*a+I*b)*a-(1/4)*(a+I*b)^2-(1/4)*(a+I*b)*(a-I*b) = 0

(14)

normal(a^2+(a+I*b)*a-(1/2)*(3*a+I*b)*a-(1/4)*(a+I*b)^2-(1/4)*(a+I*b)*(a-I*b) = 0)

0 = 0

(15)

The equivalent differentiation, this time replacing in eq__1 conjugate by abs; construct also the equivalent expression in terms of z and  conjugate(z) for verifying results

eq__3 := eval(eq__1, conjugate = abs)

abs(z)+abs(z)^2*z

(16)

eq__33 := simplify(convert(eq__3, conjugate), size)

(z*conjugate(z))^(1/2)+conjugate(z)*z^2

(17)

Since these two expressions are mathematically equal, their derivatives should be too, and the derivatives of eq__33 can be verified by eye since z and  conjugate(z) are taken as independent variables

(%diff = diff)(eq__3, z)

%diff(abs(z)+abs(z)^2*z, z) = (1/2)*conjugate(z)/abs(z)+z*conjugate(z)+abs(z)^2

(18)

(%diff = diff)(eq__33, z)

%diff((z*conjugate(z))^(1/2)+conjugate(z)*z^2, z) = (1/2)*conjugate(z)/(z*conjugate(z))^(1/2)+2*z*conjugate(z)

(19)

Eq (18) is expressed in terms of abs(z) = abs(z) while (19) is in terms of conjugate(z) = conjugate(z). Comparing as done in (14)

rhs((%diff(abs(z)+abs(z)^2*z, z) = (1/2)*conjugate(z)/abs(z)+z*conjugate(z)+abs(z)^2)-(%diff((z*conjugate(z))^(1/2)+conjugate(z)*z^2, z) = (1/2)*conjugate(z)/(z*conjugate(z))^(1/2)+2*z*conjugate(z))) = 0

(1/2)*conjugate(z)/abs(z)-z*conjugate(z)+abs(z)^2-(1/2)*conjugate(z)/(z*conjugate(z))^(1/2) = 0

(20)

`assuming`([eval((1/2)*conjugate(z)/abs(z)-z*conjugate(z)+abs(z)^2-(1/2)*conjugate(z)/(z*conjugate(z))^(1/2) = 0, z = a+I*b)], [a::real, b::real])

(1/2)*(a-I*b)/(a^2+b^2)^(1/2)-(a+I*b)*(a-I*b)+a^2+b^2-(1/2)*(a-I*b)/((a+I*b)*(a-I*b))^(1/2) = 0

(21)

simplify((1/2)*(a-I*b)/(a^2+b^2)^(1/2)-(a+I*b)*(a-I*b)+a^2+b^2-(1/2)*(a-I*b)/((a+I*b)*(a-I*b))^(1/2) = 0)

0 = 0

(22)

To mention but one not so famliar case, consider the derivative of the sign of a complex number, represented in Maple by signum(z). So our testing expression is

eq__4 := eval(eq__1, conjugate = signum)

signum(z)+z*signum(z)^2

(23)

This expression can also be rewritten in terms of z and  conjugate(z) 

eq__44 := simplify(convert(eq__4, conjugate), size)

z/(z*conjugate(z))^(1/2)+z^2/conjugate(z)

(24)

This time differentiate with respect to conjugate(z),

(%diff = diff)(eq__4, conjugate(z))

%diff(signum(z)+z*signum(z)^2, conjugate(z)) = -(1/2)*z^2/abs(z)^3-z^3*signum(z)/abs(z)^3

(25)

Here again, the differentiation of eq__44, that is expressed entirely in terms of z and  conjugate(z), can be computed by eye

(%diff = diff)(eq__44, conjugate(z))

%diff(z/(z*conjugate(z))^(1/2)+z^2/conjugate(z), conjugate(z)) = -(1/2)*z^2/(z*conjugate(z))^(3/2)-z^2/conjugate(z)^2

(26)

Eq (25) is expressed in terms of abs(z) = abs(z) while (26) is in terms of conjugate(z) = conjugate(z). Comparing as done in (14),

rhs((%diff(signum(z)+z*signum(z)^2, conjugate(z)) = -(1/2)*z^2/abs(z)^3-z^3*signum(z)/abs(z)^3)-(%diff(z/(z*conjugate(z))^(1/2)+z^2/conjugate(z), conjugate(z)) = -(1/2)*z^2/(z*conjugate(z))^(3/2)-z^2/conjugate(z)^2)) = 0

-(1/2)*z^2/abs(z)^3-z^3*signum(z)/abs(z)^3+(1/2)*z^2/(z*conjugate(z))^(3/2)+z^2/conjugate(z)^2 = 0

(27)

`assuming`([eval(-(1/2)*z^2/abs(z)^3-z^3*signum(z)/abs(z)^3+(1/2)*z^2/(z*conjugate(z))^(3/2)+z^2/conjugate(z)^2 = 0, z = a+I*b)], [a::real, b::real])

-(1/2)*(a+I*b)^2/(a^2+b^2)^(3/2)-(a+I*b)^4/(a^2+b^2)^2+(1/2)*(a+I*b)^2/((a+I*b)*(a-I*b))^(3/2)+(a+I*b)^2/(a-I*b)^2 = 0

(28)

simplify(-(1/2)*(a+I*b)^2/(a^2+b^2)^(3/2)-(a+I*b)^4/(a^2+b^2)^2+(1/2)*(a+I*b)^2/((a+I*b)*(a-I*b))^(3/2)+(a+I*b)^2/(a-I*b)^2 = 0)

0 = 0

(29)

NULL


 

Download Wirtinger_Derivatives.mw

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft

 

Download FeynmanIntegrals.mw

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft

One of the most interesting help page about the use of the Physics package is Physics,Examples. This page received some additions recently. It is also an excellent example of the File -> Export -> LaTeX capabilities under development.

Below you see the sections and subsections of this page. At the bottom, you have links to the updated PhysicsExample.mw worksheet, together with PhysicsExamples.PDF.

The PDF file has 74 pages and is obtained by going File -> Export -> LaTeX (FEL) on this worksheet to get a .tex version of it using an experimental version of Maple under development. The .tex file that results from FEL (used to get the PDF using TexShop on a Mac) has no manual editing. This illustrates new automatic line-breakingequation labels, colours, plots, and the new LaTeX translation of sophisticated mathematical physics notation used in the Physics package (command Latex in the Maplesoft Physics Updates, to be renamed as latex in the upcoming Maple release). 

In brief, this LaTeX project aims at writing entire course lessons or scientific papers directly in the Maple worksheet that combines what-you-see-is-what-you-get editing capabilities with the Maple computational engine to produce mathematical results. And from there get a LaTeX version of the work in two clicks, optionally hiding all the input (View -> Show/Hide -> Input).

PhysicsExamples.mw   PhysicsExamples.pdf

PS: MANY THANKS to all of you who provided so-valuable feedback on the new Latex here in Mapleprimes.

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft


In a recent question in Mapleprimes, a spacetime (metric) solution to Einstein's equations, from chapter 27 of the book of Exact Solutions to Einstein's equations [1] was discussed. One of the issues was about computing a tetrad for that solution [27, 37, 1] such that the corresponding Weyl scalars are in canonical form. This post illustrates how to do that, with precisely that spacetime metric solution, in two different ways: 1) automatically, all in one go, and 2) step-by-step. The step-by-step computation is useful to verify results and also to compute different forms of the tetrads or Weyl scalars. The computation below is performed using the latest version of the Maplesoft Physics Updates.

 

with(Physics)

Physics:-Version()

`The "Physics Updates" version in the MapleCloud is 851 and is the same as the version installed in this computer, created 2020, October 19, 13:47 hours Pacific Time.`

(1)

The starting point is this image of page 421 of the book of Exact Solutions to Einstein's equations, formulas (27.37)

 

Load the solution [27, 37, 1] from Maple's database of solutions to Einstein's equations

g_[[27, 37, 1]]

_______________________________________________________

 

`Systems of spacetime coordinates are:`*{X = (z, zb, r, u)}

 

`Default differentiation variables for d_, D_ and dAlembertian are:`*{X = (z, zb, r, u)}

 

`The `*`Robinson and Trautman (1962)`*` metric in coordinates `*[z, zb, r, u]

 

`Parameters: `*[P(z, zb, u), H(X)]

 

"`Comments: ` admits geodesic, shearfree, twistfree null congruence, rho=-1/r=rho_b"

 

`Resetting the signature of spacetime from `*`- - - +`*` to `*`+ + + -`*` in order to match the signature in the database of metrics`

 

_______________________________________________________

 

`Setting `*lowercaselatin_is*` letters to represent `*space*` indices`

 

Physics:-g_[mu, nu] = Matrix(%id = 18446744078276690638)

(2)

"CompactDisplay(?)"

H(X)*`will now be displayed as`*H

 

P(z, zb, u)*`will now be displayed as`*P

(3)

The assumptions on the metric's parameters are

Assume(P(z, zb, u) > 0, (H(X))::real, r >= 0)

 

The line element is as shown in the second line of the image above

g_[lineelement]

2*r^2*Physics:-d_(z)*Physics:-d_(zb)/P(z, zb, u)^2-2*Physics:-d_(r)*Physics:-d_(u)-2*H(X)*Physics:-d_(u)^2

(4)

Load Tetrads

with(Tetrads)

_______________________________________________________

 

`Setting `*lowercaselatin_ah*` letters to represent `*tetrad*` indices`

 

((`Defined as tetrad tensors `*`see <a href='http://www.maplesoft.com/support/help/search.aspx?term=Physics,tetrads`*`,' target='_new'>?Physics,tetrads`*`,</a> `*`&efr;`[a, mu]*`, `)*eta[a, b]*`, `*gamma[a, b, c]*`, `)*lambda[a, b, c]

 

((`Defined as spacetime tensors representing the NP null vectors of the tetrad formalism `*`see <a href='http://www.maplesoft.com/support/help/search.aspx?term=Physics,tetrads`*`,' target='_new'>?Physics,tetrads`*`,</a> `*l[mu]*`, `)*n[mu]*`, `*m[mu]*`, `)*conjugate(m[mu])

 

_______________________________________________________

(5)

The Petrov type of this spacetime solution is

PetrovType()

"II"

(6)

The null tetrad computed by the Maple system using a general algorithms is

Setup(tetrad = null)

e_[]

Physics:-Tetrads:-e_[a, mu] = Matrix(%id = 18446744078178770326)

(7)

 

According to the help page TransformTetrad , the canonical form of the Weyl scalars for each different Petrov type is

 

So for type II, when the tetrad is in canonical form, we expect only `&Psi;__2` and `&Psi;__3` different from 0. For the tetrad computed automatically, however, the scalars are

Weyl[scalars]

psi__0 = -P(z, zb, u)*(2*(diff(P(z, zb, u), z))*(diff(H(X), z))+P(z, zb, u)*(diff(diff(H(X), z), z)))/(r^2*(H(X)^2+1)^(1/2)), psi__1 = ((1/2)*I)*(-(diff(diff(H(X), r), z))*P(z, zb, u)^2*r+2*P(z, zb, u)^2*(diff(H(X), z))-(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*r+(diff(diff(P(z, zb, u), u), z))*r*P(z, zb, u))/(P(z, zb, u)*r^2*(H(X)^2+1)^(1/4)), psi__2 = (1/6)*((diff(diff(H(X), r), r))*r^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*(diff(H(X), r))*r-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)+2*H(X))/r^2, psi__3 = 0, psi__4 = 0

(8)

The question is, how to bring the tetrad `&efr;`[a, mu] (equation (7)) into canonical form. The plan for that is outlined in Chapter 7, by Chandrasekhar, page 388, of the book "General Relativity, an Einstein centenary survey", edited by S.W. Hawking and W.Israel. In brief, for Petrov type II, use a transformation ofClass[2] to make Psi[0] = `&Psi;__1` and `&Psi;__1` = 0, then a transformation of Class[1] making Psi[4] = 0, finally use a transformation of Class[3] making Psi[3] = 1. For an explanation of these transformations see the help page for TransformTetrad . This plan, however, is applicable if and only if the starting tetrad results in `&psi;__4` <> 0, which we see in (8) it is not the case, so we need, in addition, before applying this plan, to perform a transformation of Class[1] making `&psi;__4` <> 0.

 

In what follows, the transformations mentioned are first performed automatically, in one go, letting the computer deduce each intermediate transformation, by passing to TransformTetrad the optional argument canonicalform. Then, the same result is obtained by transforming the starting tetrad  one step at at time, arriving at the same Weyl scalars. That illustrates well both how to get the result exploiting advanced functionality but also how to verify the result performing each step, and also how to get any desired different form of the Weyl scalars.

 

Although it is possible to perform both computations, automatically and step-by-step, departing from the tetrad (7), that tetrad and the corresponding Weyl scalars (8) have radicals, making the readability of the formulas at each step less clear. Both computations, can be presented in more readable form without radicals departing from the tetrad shown in the book, that is

e_[a, mu] = (Matrix(4, 4, {(1, 1) = 0, (1, 2) = 0, (1, 3) = 0, (1, 4) = -1, (2, 1) = 0, (2, 2) = r/P(z, zb, u), (2, 3) = 0, (2, 4) = 0, (3, 1) = r/P(z, zb, u), (3, 2) = 0, (3, 3) = 0, (3, 4) = 0, (4, 1) = 0, (4, 2) = 0, (4, 3) = -1, (4, 4) = -H(X)}))

Physics:-Tetrads:-e_[a, mu] = Matrix(%id = 18446744078621688766)

(9)

"IsTetrad(?)"

`Type of tetrad: `*null

 

true

(10)

The corresponding Weyl scalars free of radicals are

"WeylScalars(?)"

psi__0 = P(z, zb, u)*(2*(diff(P(z, zb, u), z))*(diff(H(X), z))+P(z, zb, u)*(diff(diff(H(X), z), z)))/r^2, psi__1 = -(1/2)*(-(diff(diff(H(X), r), z))*P(z, zb, u)^2*r+2*P(z, zb, u)^2*(diff(H(X), z))-(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*r+(diff(diff(P(z, zb, u), u), z))*r*P(z, zb, u))/(r^2*P(z, zb, u)), psi__2 = -(1/6)*(-(diff(diff(H(X), r), r))*r^2+2*(diff(H(X), r))*r-2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))+2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)-2*H(X))/r^2, psi__3 = 0, psi__4 = 0

(11)

So set this tetrad as the starting point

"Setup(?)"

[tetrad = {(1, 4) = -1, (2, 2) = r/P(z, zb, u), (3, 1) = r/P(z, zb, u), (4, 3) = -1, (4, 4) = -H(X)}]

(12)


All the transformations performed automatically, in one go

 

To arrive in one go, automatically, to a tetrad whose Weyl scalars are in canonical form as in (31), use the optional argument canonicalform:

T__5 := TransformTetrad(canonicalform)

WeylScalars(T__5)

psi__0 = 0, psi__1 = 0, psi__2 = -(1/6)*(-(diff(diff(H(X), r), r))*r^2+2*(diff(H(X), r))*r-2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))+2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)-2*H(X))/r^2, psi__3 = 1, psi__4 = 0

(13)

Note the length of T__5

length(T__5)

58242

(14)

That length corresponds to several pages long. That happens frequently, you get Weyl scalars with a minimum of residual invariance, at the cost of a more complicated tetrad.

 

The transformations step-by-step leading to the same canonical form of the Weyl scalars

 

Step 0

 

As mentioned above, to apply the plan outlined by Chandrasekhar, the starting point needs to be a tetrad with `&Psi;__4` <> 0, not the case of (9), so in this step 0 we use a transformation of Class[1] making `&psi;__4` <> 0. This transformation introduces a complex parameter E and to get `&psi;__4` <> 0 any value of E suffices. We use E = 1:

TransformTetrad(nullrotationwithfixedl_)

Matrix(%id = 18446744078634914990)

(15)

"`T__0` := eval(?,E=1)"

Matrix(%id = 18446744078634940646)

(16)

Indeed, for this tetrad, `&Psi;__4` <> 0:

WeylScalars(T__0)[-1]

psi__4 = ((diff(diff(H(X), r), r))*r^2*P(z, zb, u)+P(z, zb, u)^3*(diff(diff(H(X), z), z))+2*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^2+2*(diff(diff(H(X), r), z))*P(z, zb, u)^2*r+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))*P(z, zb, u)+2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*r-4*P(z, zb, u)^2*(diff(H(X), z))-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)^2-2*(diff(H(X), r))*P(z, zb, u)*r-2*(diff(diff(P(z, zb, u), u), z))*r*P(z, zb, u)+2*H(X)*P(z, zb, u))/(r^2*P(z, zb, u))

(17)

Step 1

Next is a transformation of Class__2 to make `&Psi;__0` = 0, that in the case of Petrov type II also implies on `&Psi;__1` = 0.According to the the help page TransformTetrad , this transformation introduces a parameter B that, according to the plan outlined by Chandrasekhar in Chapter 7 page 388, is one of the two identical roots (out of the four roots) of the principalpolynomial. To see the principal polynomial, or, directly, its roots you can use the PetrovType  command:

PetrovType(principalroots = 'R')

"II"

(18)

The first two are the same and equal to -1

R[1 .. 2]

[-1, -1]

(19)

So the transformed tetrad T__1 is

T__1 := eval(TransformTetrad(T__0, nullrotationwithfixedn_), B = -1)

Matrix(%id = 18446744078641721462)

(20)

Check this result and the corresponding Weyl scalars to verify that we now have `&Psi;__0` = 0 and `&Psi;__1` = 0

IsTetrad(T__1)

`Type of tetrad: `*null

 

true

(21)

WeylScalars(T__1)[1 .. 2]

psi__0 = 0, psi__1 = 0

(22)

Step 2

Next is a transformation of Class__1 that makes `&Psi;__4` = 0. This transformation introduces a parameter E, that according to Chandrasekhar's plan can be taken equal to one of the roots of Weyl scalar `&Psi;__4`that corresponds to the transformed tetrad. So we need to proceed in three steps:

a. 

transform the tetrad introducing a parameter E in the tetrad's components

b. 

compute the Weyl scalars for that transformed tetrad

c. 

take `&Psi;__4` = 0 and solve for E

d. 

apply the resulting value of E to the transformed tetrad obtained in step a.

 

a.Transform the tetrad and for simplicity take E real

T__2 := eval(TransformTetrad(T__1, nullrotationwithfixedl_), conjugate(E) = E)

Matrix(%id = 18446744078624751238)

(23)

"IsTetrad(?)"

`Type of tetrad: `*null

 

true

(24)

b. Compute `&Psi;__4` for this tetrad

simplify(WeylScalars(T__2)[-1])

psi__4 = (r^2*P(z, zb, u)*(E-1)^2*(diff(diff(H(X), r), r))-2*r*P(z, zb, u)^2*(E-1)*(diff(diff(H(X), r), z))+P(z, zb, u)^3*(diff(diff(H(X), z), z))-2*P(z, zb, u)^2*(E-1)^2*(diff(diff(P(z, zb, u), z), zb))+2*r*P(z, zb, u)*(E-1)*(diff(diff(P(z, zb, u), u), z))-2*r*P(z, zb, u)*(E-1)^2*(diff(H(X), r))+4*P(z, zb, u)^2*(E+(1/2)*(diff(P(z, zb, u), z))-1)*(diff(H(X), z))+2*((P(z, zb, u)*(E-1)*(diff(P(z, zb, u), zb))-(diff(P(z, zb, u), u))*r)*(diff(P(z, zb, u), z))+H(X)*P(z, zb, u)*(E-1))*(E-1))/(r^2*P(z, zb, u))

(25)

c. Solve `&Psi;__4` = 0 discarding the case E = 0 which implies on no transformation

simplify(solve({rhs(psi__4 = (r^2*P(z, zb, u)*(E-1)^2*(diff(diff(H(X), r), r))-2*r*P(z, zb, u)^2*(E-1)*(diff(diff(H(X), r), z))+P(z, zb, u)^3*(diff(diff(H(X), z), z))-2*P(z, zb, u)^2*(E-1)^2*(diff(diff(P(z, zb, u), z), zb))+2*r*P(z, zb, u)*(E-1)*(diff(diff(P(z, zb, u), u), z))-2*r*P(z, zb, u)*(E-1)^2*(diff(H(X), r))+4*P(z, zb, u)^2*(E+(1/2)*(diff(P(z, zb, u), z))-1)*(diff(H(X), z))+2*((P(z, zb, u)*(E-1)*(diff(P(z, zb, u), zb))-(diff(P(z, zb, u), u))*r)*(diff(P(z, zb, u), z))+H(X)*P(z, zb, u)*(E-1))*(E-1))/(r^2*P(z, zb, u))) = 0, E <> 0}, {E}, explicit)[1])

{E = ((diff(diff(H(X), r), r))*r^2*P(z, zb, u)+(diff(diff(H(X), r), z))*P(z, zb, u)^2*r-2*(diff(H(X), r))*P(z, zb, u)*r-(diff(diff(P(z, zb, u), u), z))*r*P(z, zb, u)+(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*r-2*P(z, zb, u)^2*(diff(H(X), z))-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))*P(z, zb, u)+2*H(X)*P(z, zb, u)+(-P(z, zb, u)^4*((diff(diff(H(X), r), r))*r^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*(diff(H(X), r))*r-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)+2*H(X))*(diff(diff(H(X), z), z))+P(z, zb, u)^4*(diff(diff(H(X), r), z))^2*r^2+(-2*r^2*(diff(diff(P(z, zb, u), u), z))*P(z, zb, u)^3+2*r^2*P(z, zb, u)^2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))-4*r*(diff(H(X), z))*P(z, zb, u)^4)*(diff(diff(H(X), r), z))-2*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(H(X), r), r))*r^2+P(z, zb, u)^2*(diff(diff(P(z, zb, u), u), z))^2*r^2+(-2*r^2*P(z, zb, u)*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))+4*r*(diff(H(X), z))*P(z, zb, u)^3)*(diff(diff(P(z, zb, u), u), z))+4*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(P(z, zb, u), z), zb))+4*(diff(H(X), z))^2*P(z, zb, u)^4+4*P(z, zb, u)^2*(diff(P(z, zb, u), z))*((diff(H(X), r))*P(z, zb, u)*r-(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))*P(z, zb, u)-(diff(P(z, zb, u), u))*r-H(X)*P(z, zb, u))*(diff(H(X), z))+(diff(P(z, zb, u), u))^2*(diff(P(z, zb, u), z))^2*r^2)^(1/2))/(P(z, zb, u)*((diff(diff(H(X), r), r))*r^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*(diff(H(X), r))*r-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)+2*H(X)))}

(26)

d. Apply this result to the tetrad (23). In doing so, do not display the result, just measure its length (corresponds to two+ pages)

T__3 := simplify(eval(T__2, {E = ((diff(diff(H(X), r), r))*r^2*P(z, zb, u)+(diff(diff(H(X), r), z))*P(z, zb, u)^2*r-2*(diff(H(X), r))*P(z, zb, u)*r-(diff(diff(P(z, zb, u), u), z))*r*P(z, zb, u)+(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*r-2*P(z, zb, u)^2*(diff(H(X), z))-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))*P(z, zb, u)+2*H(X)*P(z, zb, u)+(-P(z, zb, u)^4*((diff(diff(H(X), r), r))*r^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*(diff(H(X), r))*r-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)+2*H(X))*(diff(diff(H(X), z), z))+P(z, zb, u)^4*(diff(diff(H(X), r), z))^2*r^2+(-2*r^2*(diff(diff(P(z, zb, u), u), z))*P(z, zb, u)^3+2*r^2*P(z, zb, u)^2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))-4*r*(diff(H(X), z))*P(z, zb, u)^4)*(diff(diff(H(X), r), z))-2*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(H(X), r), r))*r^2+P(z, zb, u)^2*(diff(diff(P(z, zb, u), u), z))^2*r^2+(-2*r^2*P(z, zb, u)*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))+4*r*(diff(H(X), z))*P(z, zb, u)^3)*(diff(diff(P(z, zb, u), u), z))+4*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(P(z, zb, u), z), zb))+4*(diff(H(X), z))^2*P(z, zb, u)^4+4*P(z, zb, u)^2*(diff(P(z, zb, u), z))*((diff(H(X), r))*P(z, zb, u)*r-(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))*P(z, zb, u)-(diff(P(z, zb, u), u))*r-H(X)*P(z, zb, u))*(diff(H(X), z))+(diff(P(z, zb, u), u))^2*(diff(P(z, zb, u), z))^2*r^2)^(1/2))/(P(z, zb, u)*((diff(diff(H(X), r), r))*r^2+2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*(diff(H(X), r))*r-2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)+2*H(X)))}[1]))

length(T__3)

12589

(27)

Check the scalars, we expect `&Psi;__0` = `&Psi;__1` and `&Psi;__1` = `&Psi;__4` and `&Psi;__4` = 0

WeylScalars(T__3); %[1 .. 2], %[-1]

psi__0 = 0, psi__1 = 0, psi__4 = 0

(28)

Step 3

Use a transformation of Class[3] making Psi[3] = 1. Such a transformation changes Psi[3]^` '` = A*exp(-I*Omega)*Psi[3], where we need to take A*exp(-I*Omega) = 1/`&Psi;__3`, and without loss of generality we can take Omega = 0.

Check first the value of `&Psi;__3` in the last tetrad computed

WeylScalars(T__3)[4]

psi__3 = (1/2)*(-2*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(H(X), r), r))*r^2-P(z, zb, u)^4*(diff(diff(H(X), z), z))*(diff(diff(H(X), r), r))*r^2+P(z, zb, u)^4*(diff(diff(H(X), r), z))^2*r^2+4*(diff(P(z, zb, u), z))*(diff(H(X), r))*(diff(H(X), z))*P(z, zb, u)^3*r+2*(diff(H(X), r))*P(z, zb, u)^4*(diff(diff(H(X), z), z))*r-4*(diff(P(z, zb, u), zb))*(diff(P(z, zb, u), z))^2*(diff(H(X), z))*P(z, zb, u)^3+4*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(P(z, zb, u), z), zb))-4*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(H(X), r), z))*r+2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*P(z, zb, u)^2*(diff(diff(H(X), r), z))*r^2-2*(diff(P(z, zb, u), zb))*(diff(P(z, zb, u), z))*P(z, zb, u)^4*(diff(diff(H(X), z), z))+2*P(z, zb, u)^5*(diff(diff(H(X), z), z))*(diff(diff(P(z, zb, u), z), zb))-2*P(z, zb, u)^3*(diff(diff(P(z, zb, u), u), z))*(diff(diff(H(X), r), z))*r^2+4*(diff(H(X), z))^2*P(z, zb, u)^4-4*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^2*r-4*H(X)*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3+4*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(P(z, zb, u), u), z))*r+(diff(P(z, zb, u), u))^2*(diff(P(z, zb, u), z))^2*r^2-2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*P(z, zb, u)*(diff(diff(P(z, zb, u), u), z))*r^2-2*H(X)*P(z, zb, u)^4*(diff(diff(H(X), z), z))+P(z, zb, u)^2*(diff(diff(P(z, zb, u), u), z))^2*r^2)^(1/2)/(r^2*P(z, zb, u))

(29)

So, the transformed tetrad T__4 to which corresponds Weyl scalars in canonical form, with `&Psi;__0` = `&Psi;__1` and `&Psi;__1` = `&Psi;__4` and `&Psi;__4` = 0 and `&Psi;__3` = 1, is

T__4 := simplify(eval(TransformTetrad(T__3, boostsn_l_plane), A = 1/rhs(psi__3 = (1/2)*(-2*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(H(X), r), r))*r^2-P(z, zb, u)^4*(diff(diff(H(X), z), z))*(diff(diff(H(X), r), r))*r^2+P(z, zb, u)^4*(diff(diff(H(X), r), z))^2*r^2+4*(diff(P(z, zb, u), z))*(diff(H(X), r))*(diff(H(X), z))*P(z, zb, u)^3*r+2*(diff(H(X), r))*P(z, zb, u)^4*(diff(diff(H(X), z), z))*r-4*(diff(P(z, zb, u), zb))*(diff(P(z, zb, u), z))^2*(diff(H(X), z))*P(z, zb, u)^3+4*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(P(z, zb, u), z), zb))-4*(diff(H(X), z))*P(z, zb, u)^4*(diff(diff(H(X), r), z))*r+2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*P(z, zb, u)^2*(diff(diff(H(X), r), z))*r^2-2*(diff(P(z, zb, u), zb))*(diff(P(z, zb, u), z))*P(z, zb, u)^4*(diff(diff(H(X), z), z))+2*P(z, zb, u)^5*(diff(diff(H(X), z), z))*(diff(diff(P(z, zb, u), z), zb))-2*P(z, zb, u)^3*(diff(diff(P(z, zb, u), u), z))*(diff(diff(H(X), r), z))*r^2+4*(diff(H(X), z))^2*P(z, zb, u)^4-4*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^2*r-4*H(X)*(diff(P(z, zb, u), z))*(diff(H(X), z))*P(z, zb, u)^3+4*(diff(H(X), z))*P(z, zb, u)^3*(diff(diff(P(z, zb, u), u), z))*r+(diff(P(z, zb, u), u))^2*(diff(P(z, zb, u), z))^2*r^2-2*(diff(P(z, zb, u), u))*(diff(P(z, zb, u), z))*P(z, zb, u)*(diff(diff(P(z, zb, u), u), z))*r^2-2*H(X)*P(z, zb, u)^4*(diff(diff(H(X), z), z))+P(z, zb, u)^2*(diff(diff(P(z, zb, u), u), z))^2*r^2)^(1/2)/(r^2*P(z, zb, u)))))

IsTetrad(T__4)

`Type of tetrad: `*null

 

true

(30)

WeylScalars(T__4)

psi__0 = 0, psi__1 = 0, psi__2 = -(1/6)*(-(diff(diff(H(X), r), r))*r^2+2*(diff(H(X), r))*r+2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)-2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))-2*H(X))/r^2, psi__3 = 1, psi__4 = 0

(31)

These are the same scalars computed in one go in (13)

psi__0 = 0, psi__1 = 0, psi__2 = -(1/6)*(-(diff(diff(H(X), r), r))*r^2+2*(diff(H(X), r))*r-2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))+2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)-2*H(X))/r^2, psi__3 = 1, psi__4 = 0

psi__0 = 0, psi__1 = 0, psi__2 = -(1/6)*(-(diff(diff(H(X), r), r))*r^2+2*(diff(H(X), r))*r-2*(diff(P(z, zb, u), z))*(diff(P(z, zb, u), zb))+2*(diff(diff(P(z, zb, u), z), zb))*P(z, zb, u)-2*H(X))/r^2, psi__3 = 1, psi__4 = 0

(32)

``


 

Download The_metric_[27_37_1]_in_canonical_form.mw

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft

Hi,
Some people using the Windows platform have had problems installing MapleCloud packages, including the Maplesoft Physics Updates. This problem does not happen in Macintosh or Linux/Unix, also does not happen with all Windows computers but with some of them, and is not a problem of the MapleCloud packages themselves, but a problem of the installer of packages.

I understand that a solution to this problem will be presented within an upcoming Maple dot release.

Meantime, there is a solution by installing a helper library; after that, MapleCloud packages install without problems in all Windows machines. So whoever is having trouble installing MapleCloud packages in Windows and prefers not to wait until that dot release, instead wants to install this helper library, please email me at physics@maplesoft.com

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft

1 2 3 4 5 6 7 Last Page 1 of 15