Which convex hull algorithm is this package based on?

[WuSiren]

Hi, everyone! May I ask which convex hull algorithm is this package based on (Are there any references?)? And what are the differences between this package and the convex hull algorithm introduced here? Which one is more efficient?

Thanks in advance!

[blegat]

You have a list of different libraries you can choose from https://juliapolyhedra.github.io/. If you use the default implementation in Polyhedra.jl, it depends whether you removing redundancy or computing the H-representation, it would help to have a code example of what you are doing.

[WuSiren]

Oh, I see, thanks a lot! I'm just curious about what the underlying algorithm of this package is. Now I see it's the Quickhull. Thanks! :)

Qhull implements the Quickhull algorithm for computing the convex hull.

[blegat]

Ah yes, I missed that the issue was opened on this package and not Polyhedra.jl, yest it's this one http://www.qhull.org/

[WuSiren]

Thanks to you and these valuable packages! 👍

[WuSiren]

Hey, @blegat , I'm here again. I would like to ask whether QHull.jl is allowed to construct a convex hull for a given set of points that covers more than $(1-\epsilon)$ of the points? Or is there any other typical method to do so? Thanks!

[blegat]

Not aware of any library doing that. I would probably first compute the convex hull of all the points and then look at the sensitivity, ideally with a linear programming approach.

[WuSiren]

OK. Many thanks. 🤝