AlgebraicCurveOrthogonalPolynomials.jl

Author JuliaApproximation
Popularity
1 Star
Updated Last
5 Months Ago
Started In
February 2020

AlgebraicCurveOrthogonalPolynomials.jl

A Julia package for orthogonal polynomials on algebraic curves

This package contains ongoing research on orthogonal polynomials on algebraic curves. That is, given an algebraic curve in 2D defined by S = {(x,y) : p(x,y) = 0} for some bivariate polynomial p we wish to construct polynomials orthogonal with respect to an inner product supported on a subset Ω ⊆ S. Note to avoid degenerecies these polynomials need to be thought of as polynomials modulo the vanishing ideal I(S) associated to S. The general theory and construction is not at yet possible, so we outline some specific cases that we have implemented below.

This is funded by a Leverhulme Trust Research Project Grant on "Constructive approximation on algebraic curves and surfaces".

Arc

We can construct orthogonal polynomials on an arc, that is, Ω = {(cos(θ), sin(θ)) : a ≤ θ ≤ b}, which is a subset of the circle {(x,y) : x^2 + y^2 = 1}. We parameterise points on the circle by angle, using a special type CircleCoordinate(θ):

julia> CircleCoordinate(0.1)
2-element CircleCoordinate{Float64} with indices SOneTo(2):
 0.9950041652780258
 0.09983341664682815

For now we only support the half circle y ≥ 0 with the weight y^a, which we construct via UltrasphericalArc(a), which is implemented in the framework of ContinuumArrays.jl:

julia> P = UltrasphericalArc() # uniform weight on the arc
UltrasphericalArc(0.0)

julia> P[CircleCoordinate(0.1),1:5] # first 5 polynomials
5-element Array{Float64,1}:
 1.0
 0.9950041652780258
 0.7024490016371341
 2.030105652576658
 0.06160390817639964

Note there are two (and only two) degree-d polynomials apart from d = 1. This is accessible as the columns of P are blocked a la BlockArrays.jl:

julia> P[CircleCoordinate(0.1), Block.(1:3)]
3-blocked 5-element PseudoBlockArray{Float64,1,Array{Float64,1},Tuple{BlockedUnitRange{StepRange{Int64,Int64}}}}:
 1.0                
 ───────────────────
 0.9950041652780258 
 0.7024490016371341 
 ───────────────────
 2.030105652576658  
 0.06160390817639964