Re: [P-1788]: Re objective == infinity
On Sep 28 2012, Ralph Baker Kearfott wrote:
Does this mean we should revisit our decision that
intervals are subsets of the set of real numbers?
That would be problematical. However, I don't
think proceeding with subsets of real numbers only
is such a huge problem. It would probably be easier
to resolve the issue without revisiting the
issue of extended reals.
Er, I hope that you don't mean what I think you mean!
Once you get beyond simple intervals, subsets get both
mathematically hard and computationally evil. For
example, if you have a subset [0.0,3.0),(3.0,9.0] and
transform it so that the location corresponding to 3.0
is non-representable in any normal finite arithmetic,
is that equivalent to the subset [0.0,9.0]?
Regards,
Nick Maclaren.