The general form of the coordinates is:
coords = r $\{$$\cos\theta$ $\sin\phi$, $\sin\theta$ $\sin\phi$, $\cos\phi$}
I've considered the radius $r=1$. Now varying the angles $\theta$ and $\phi$, I can generate equidistant points on the surface of the sphere.
But, I want those points to represent the vertices of a regular polyhedron (not in a true sense always; for example, I will consider a triangular bipyramid for $n=5$) for $n\ge4$, where $n$ is the number of points so that the vertices lie on the surface of the sphere. In this way, probably (please correct me if I am wrong) I can have polyhedron vertices for any $n\ge4$.
Can I have any formula to perform this task?
I don't find your question very clear.
For (r, θ, ϕ) = (radial, azimuthal, polar) (!)
So why not write just (x,y,z)?
Maybe you mean to use spherical coordinates
instead?
You say you can already generate equidistant points, by varying theta and phi. If you vary theta by 360/i degrees where i is an integer in the range [1,360]. (Or alternatively phi), you have your equidistant points already, haven't you?
I would think just calculating the Euclidian distance between 2 points on the sphere (I would use cartesian coordinates instead to this end) is enough as the size of the sphere is fixed. d = sqrt(xx + yy + z*z)
Because you asked for a formula, you should be able to derive it from here: http://mathworld.wolfram.com/SphericalCoordinates.html