[R] reasonable theory?

[James Cloos]

>>>>> "SG" == Spencer Graves <spencer.graves at structuremonitoring.com> writes:

SG> If you know nothing about the black box except that its domain is
SG> bounded, then I would random sample uniformly from the domain.  If the
SG> function is monotonically increasing in all variables, then you only
SG> need to test the two extreme points.  If you know other things, you
SG> may be able to use the other logic.

So I was overthinking it, then.

As it turns out, for most of the functions in question, the convex hull
of the range, as determined by uniform samples w/in the domain, was only
slightly bigger than the convex hull of the full set of extreme points
of the domain.

And for my needs, then, the latter is a good-enough approximation of the
former after all.

Thanks!

-JimC
-- 
James Cloos <cloos at jhcloos.com>         OpenPGP: 1024D/ED7DAEA6