From: Johannes H Andersen Subject: Re: Polynomial basis functions on the interval [0,1] Date: Sat, 25 Mar 2000 23:09:50 +0000 Newsgroups: sci.math Summary: [missing] Les Wright wrote: > > Hiram Berry wrote: > > > Thankyou for the answer and reference, Johannes. That's just the type of > > answer I was looking for; in the meantime, while staring for a while at the > > sequence of polynomials previously found, I did finally see the pattern for > > recursively generating the particular sequence I posted previously: > > > > call the integer coefficient (excluding the sqrt part) of the m-order term > > in the n-order polynomial Pn previously defined, "T(n,m)", then > > > > T(0,0) = 1 > > T(n,0) = -1 * T(n-1,0), n > 0 > > T(n,m) = (-(n+m)/(n-m)) * T(n-1,m) , 0 < m < n > > T(n,n) = ((4*n-2)/n) * T(n-1,n-1) > > > > A much cleaner and faster algorithm than what I was doing before! > > Congratulations! > > Orthogonal polynomials intrigue me because of their applications (I am > interested in classical quantum mechanics, the Schrodinger equation and all > that), but I have never completely understood them or how do generate them with > recursion relations and generating functions. I eyeballed your question for a > time trying to see the pattern, and I am glad you found it. > > Orthogonal polynomials are often associated with particular ODEs (I am thinking > here of Legendre, Laguerre, and Hermite polynomials). Any idea which ODE, if > any, your equations are associated with? Notice that legendre Pol are defined in [-1,1] . This interval can also be described as sin(theta) , theta from south-pole to north-pole on a sphere. These polynomials are a subset of Spherical Harmonics corresponding to wavenumber zero round equator. The general spherical harmonics are : Y_m,n (theta,lambda) = Pol_m,n (sin(theta)) exp(i m lambda) . The wave 0 Legendre Pol is: P(x) = Pol_0,n (x) , x in [-1,1] , x=sin(theta) is redundant. Y_m,n are eigenfunctions of the Laplacian on a sphere: Del^2 Y_m,n () = n(n+1) Y_m,n () Johannes