The package is meant to do a few standard geometric computations accurately and quickly, using floating-point numbers. One of those computations is finding the convex hull of a set of points. As an example, we pick 100 points in the plane, where the - and -coordinate are independent and standard normally distributed.
To obtain the convex hull, we use the command.
| (1.1) |
What is returned is a list of numbers, to be understood as indices into , where the number represents the point in the input. Taking these points in the order given in the return value, we walk along the convex hull. We can plot the result as follows.
In higher-dimensional spaces, a convex hull cannot be represented as a sequence of points. In such cases, the command returns a list of the facets of the convex hull.
| (1.3) |
| |
In higher dimensions, the command works the same way, but you can't visualize the result as easily anymore, and the number of facets grows quickly.