Digi Area Group - Math Tools for Professionals
   Maple and Mathematica packages - math tools for professionals

atlas™ - modern differential geometry for Maple™

> Features List & Examples
> Template Worksheets
> Screenshots
> Documentation & Downloads
> License & Pricing
> Buy Online

 
 
 
 
Google

Features List & Examples  |  Introduction  |  Dimension  |  Indexing  |  Forms  |  Metric  |  atlasWizard - Maplet™

Ricci - flat warped product

Copyright © 2004-2008  DigiArea Group . All rights reserved.

Description:

  • Einstein manifolds (manifolds with constant Ricci curvature) are riemannian manifolds with metric tensor field g  and Ricci tensor field r = lambda*g  where lambda = const .(see Arthur L. Besse. "Einstein Manifolds"  Springer-Verlag). Thus for Ricci flat manifolds we have lambda = 0  and   r = 0 .
  • Warped products is a simple class of riemannian submersions which is defined as follows. Let {B, g[B]}  be riemannian manifold B  (base space) with metric g[B]  and {F, g[F]}  be riemannian manifold F  (fiber space) with metric g[F] . Then riemannian manifold {B*cross*F, g[B]+f*g[F]}  (total space) where f = f(b)  is positive function on B  is warped product with warped function f .
  • In this worksheet we deal with warped product with 2-dimensional base B = R^2  and metric:
    g = `&.`(d(r),d(r))/(1-r^(1-p))+4*(1-r^(1-p))*`&.`(d(theta),d(theta))/((p-1)^2)+r^2*g[F]   where g[F]  is complete metric on p-dimensional Einstein manifold F  with Ricci constant lambda[F] = p-1  and p = dim(F) . We take p-dimensional sphere {S^p, g[can]}  as the fiber space {F, g[F]} .

R^2*cross*S^p

restart:
with(atlas):


Total space - M


Domain(M);

M

Constants:
Constants(Lambda);

{Lambda, -I, Catalan, Pi, I, _Z}

Vector fields:
Vectors(E[i],X,Y,Z);

{E[i], X, Y, Z}

Differential p-forms:
Forms(e[j]=1);

{e[j]}

p-Sphere dimension (change it here):
p:=2;

p := 2

Coframe 1-forms:
Coframe(e[1]=d(rho),e[2]=d(theta),seq(e[i]=d(x[i-2]),i=3..p+2));

[e[1] = d(rho), e[2] = d(theta), e[3] = d(x[1]), e[4] = d(x[2])]

Frame vector fields:
Frame(E[i]);

[E[1] = Diff(``,rho), E[2] = Diff(``,theta), E[3] = Diff(``,x[1]), E[4] = Diff(``,x[2])]

Metric tensor fie ld:
Metric( g=1/(1-rho^(1-p))*d(rho)&.d(rho)+4*(1-rho^(1-p))/(p-1)^2*d(theta)&.d(theta)+rho^2*4*add(d(x[i])&.d(x[i]),i=1..p)/(1+add(x[i]*x[i],i=1..p))^2);

g = 1/(1-1/rho)*`&.`(e[1],e[1])+4*(1-1/rho)*`&.`(e[2],e[2])+4*rho^2*(`&.`(e[3],e[3])+`&.`(e[4],e[4]))/(1+x[1]^2+x[2]^2)^2

Connection 1-forms:
Connection(omega);

omega[i,j]

Curvature 2-forms:
Curvature(Omega);

Omega[i,j]

Curvature tensor field:
Riemann(R);

R = 4/rho^3*`&.`(`&^`(e[1],e[2]),`&^`(e[1],e[2]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[3]),`&^`(e[1],e[3]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[4]),`&^`(e[1],e[4]))-8/rho^2/(1+x[1]...
R = 4/rho^3*`&.`(`&^`(e[1],e[2]),`&^`(e[1],e[2]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[3]),`&^`(e[1],e[3]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[4]),`&^`(e[1],e[4]))-8/rho^2/(1+x[1]...
R = 4/rho^3*`&.`(`&^`(e[1],e[2]),`&^`(e[1],e[2]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[3]),`&^`(e[1],e[3]))-2/(1+x[1]^2+x[2]^2)^2/(rho-1)*`&.`(`&^`(e[1],e[4]),`&^`(e[1],e[4]))-8/rho^2/(1+x[1]...

Verify that total space is Ricci flat:
Ricci(r);

r = 0

Base space - B

Declare base domain:
Domain(B);

B

Declare forms:
Forms(u[i]=1);

{e[j], u[i]}

Declare vectors:
Vectors(U[k]);

{U[k]}

Declare coframe
Coframe(u[1]=d(zeta),u[2]=d(xi));

[u[1] = d(zeta), u[2] = d(xi)]

Frame declaration:
Frame(U[k]);

[U[1] = Diff(``,zeta), U[2] = Diff(``,xi)]

Let us define metric on the base:
Metric(G=1/(1-zeta^(1-p))*d(zeta)&.d(zeta)+4*(1-zeta^(1-p))/(p-1)^2*d(xi)&.d(xi));

G = 1/(1-1/zeta)*`&.`(u[1],u[1])+4*(1-1/zeta)*`&.`(u[2],u[2])

Submersion definition

Let us define submersion   pi : proc (M) options operator, arrow; B end proc  such that   zeta = rho, xi = theta :
Mapping(pi,M,B,
        zeta=rho,
        xi=theta);

pi

M.`--->`.B

Who(pi);
pi: mapping


TABLE([equations = [zeta = rho, xi = theta], frame = {E[4] = 0, E[3] = 0, E[1] = U[1], E[2] = U[2]}, coframe = {u[1] = e[1], u[2] = e[2]}, natural = {Diff(``,rho) = Diff(``,zeta), Diff(``,theta) = Diff...
TABLE([equations = [zeta = rho, xi = theta], frame = {E[4] = 0, E[3] = 0, E[1] = U[1], E[2] = U[2]}, coframe = {u[1] = e[1], u[2] = e[2]}, natural = {Diff(``,rho) = Diff(``,zeta), Diff(``,theta) = Diff...
TABLE([equations = [zeta = rho, xi = theta], frame = {E[4] = 0, E[3] = 0, E[1] = U[1], E[2] = U[2]}, coframe = {u[1] = e[1], u[2] = e[2]}, natural = {Diff(``,rho) = Diff(``,zeta), Diff(``,theta) = Diff...
TABLE([equations = [zeta = rho, xi = theta], frame = {E[4] = 0, E[3] = 0, E[1] = U[1], E[2] = U[2]}, coframe = {u[1] = e[1], u[2] = e[2]}, natural = {Diff(``,rho) = Diff(``,zeta), Diff(``,theta) = Diff...

Projectors of the submersion

Now we can calculate vertical projector V  and horizontal projector H :
P:=Projectors(pi);

P := TABLE([horizontal = `&.`(e[1],E[1])+`&.`(e[2],E[2]), vertical = `&.`(e[3],E[3])+`&.`(e[4],E[4])])

V:=P[vertical];

V := `&.`(e[3],E[3])+`&.`(e[4],E[4])

H:=P[horizontal];

H := `&.`(e[1],E[1])+`&.`(e[2],E[2])

Thus vertical and horizontal projections of arbitrary vector X are:

'iota[X](V)'=iota[X](V);
'iota[X](H)'=iota[X](H);

iota[X](V) = iota[X](e[3])*E[3]+iota[X](e[4])*E[4]

iota[X](H) = iota[X](e[1])*E[1]+iota[X](e[2])*E[2]

Invariants T and A of the submersion:

Let us calculate invariants of the submersion:

Inv:=Invariants(pi);

Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...
Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...
Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...
Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...
Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...
Inv := TABLE([integrabilityObstruction = 0, A = 0, T = TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)...

So, submersion invariant A[X](Y) = H*D[H*X](V*Y)+V*D[H*X](H*Y)  is equal to zero. Thus obstruction against integrability of the horizontal distribution is equal to zero. It is obvious that the submersion is a riemannian one but we can verify it directly. To do this we "rise" G metric into total space using restriction operator `&/`:

G &/ pi;

4*(rho-1)/rho*`&.`(e[2],e[2])+rho/(rho-1)*`&.`(e[1],e[1])

We obtain the horizontal part of g metric.

For tensor field T[X](Y) = H*D[V*X](V*Y)+V*D[V*X](H*Y)  we have:

T:=eval(Inv['T']);

T := TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]]])
T := TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]]])
T := TABLE(zero,[(3, 1) = [1/rho*E[3]], (4, 4) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]], (4, 1) = [1/rho*E[4]], (3, 3) = [-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*E[1]]])

To construct the T - tensor:

TT:=add(add(`&.`(e[i],e[j],T[i,j]),j=1..p+2),i=1..p+2);

TT := 1/rho*`&.`(e[3],e[1],E[3])-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*`&.`(e[3],e[3],E[1])+1/rho*`&.`(e[4],e[1],E[4])-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*`&.`(e[4],e[4],E[1])

For vectors X and Y:

'T[X](Y)'=iota[X,Y](TT);

T[X](Y) = 1/rho*iota[Y](e[3])*iota[X](e[1])*E[3]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*iota[Y](e[3])*iota[X](e[3])*E[1]+1/rho*iota[Y](e[4])*iota[X](e[1])*E[4]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*iota[Y](e[4])*iota[X...
T[X](Y) = 1/rho*iota[Y](e[3])*iota[X](e[1])*E[3]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*iota[Y](e[3])*iota[X](e[3])*E[1]+1/rho*iota[Y](e[4])*iota[X](e[1])*E[4]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*iota[Y](e[4])*iota[X...

Jump to total manifold:

Domain(M);

M

Now for coordinate representation of T we obtain:

'iota[E[i],E[j]](T)'=iota[E[i],E[j]](TT);

iota[E[i],E[j]](T) = 1/rho*delta[3,j]*delta[1,i]*E[3]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*delta[3,j]*delta[3,i]*E[1]+1/rho*delta[4,j]*delta[1,i]*E[4]-4/(1+x[1]^2+x[2]^2)^2*(rho-1)*delta[4,j]*delta[4,i]*E[1]

For mean curvature vector field N = trace[V*g](T)  where V*g  is vertical projection of the metric tensor g  we obtain:

N:=eval(Inv[meanCurvature]);

N := -2/rho^2*(rho-1)*E[1]

But for warped product we have N = -p*grad(sqrt(f))/sqrt(f)  , where f = rho^2  is warped function. Let us verify that:

'N'=-p/rho*grad(rho);

N = -2/rho^2*(rho-1)*E[1]