mvmesh (version 1.6)

UnitSphere: Define a mesh on a unit ball in n-dimensions

Description

Subdivide the unit ball or sphere into approximately equal simplices in arbitrary dimenions.

Usage

UnitSphere(n, k, method = "dyadic", p = 2, positive.only = FALSE)
UnitSphereEdgewise(n, k, p, positive.only)
UnitSphereDyadic(n, k, start = "diamond", p, positive.only)
UnitBall( n, k, method="dyadic", p=2, positive.only=FALSE )
LpNorm(x, p)

Arguments

n

Dimension of the space; the unit sphere is an (n-1) dimensional manifold

k

Number of subdivisions

method

"dyadic" or "edgewise": the former recursively subdivides the sphere to get a more uniform grid; the latter uses a faster method using one edgewise subdivision.

p

Power used in the l^p norm; p=2 is the Euclidean norm

positive.only

TRUE means restrict to the positive orthant; FALSE gives the full ball

start

starting shape: "diamond" or "icosahedron"

x

Matrix of points in n-dimensions; each column is a point in R^n

Value

an object of class "mvmesh" as described in mvmesh.

Details

UnitSphere computes a hyperspherical triangle approximation to the unit sphere. It calls either UnitSphereDyadic or UnitSphereEdgewise based on 'method'. Both work by subdividing the first orthant, and then rotating that subdivision around to other orthants. This is important for some uses of these functions; it guarantees that all vertices of a simplex are in a single orthant. Note that 'k' has a different meaning for the different methods. When method="dyadic", k specifies the number of dyadic subdivisions. When method="edgewise", k specifies the number of subdivisions as in UnitSimplex, which is then projected outward to the unit sphere. So when n=2, a dyadic subdivision with k=2 will result in 16 edges, whereas an edgewise subdivions with k=2 results in 8 edges.

UnitBall computes an approximate simplicial approximation to the unit ball. Specifically, it generates cones with one vertex at the origin and the other vertices on the surface of the unit sphere; these later vertices are from UnitSphere. If k is large, these cones will be very narrow/thin.

Examples

# NOT RUN {
UnitSphere( n=2, k=2, method="edgewise", positive.only=TRUE )
UnitSphere( n=2, k=2, method="edgewise" )

UnitSphere( n=3, k=2, method="edgewise", positive.only=TRUE )
UnitSphere( n=3, k=2, method="edgewise" )

UnitBall( n=2, k=2, method="edgewise", positive.only=TRUE )
UnitBall( n=2, k=2, method="edgewise" )

UnitSphere( n=3, k=2, method="dyadic", positive.only=TRUE )
UnitSphere( n=3, k=2, method="dyadic" )

UnitBall( n=3, k=2, method="dyadic", positive.only=TRUE )
UnitBall( n=3, k=2, method="dyadic" )

UnitSphere(  n=3, k=2 )
UnitBall( n=3, k=2 )

x <- c(3,-1,2)
LpNorm( x, p=2 )

# }
# NOT RUN {
plot( UnitSphere(  n=3, k=2 ), show.label=TRUE )
plot( UnitBall( n=3, k=2 ) )

# }