5

I have the intuition that the minimal point (in the sense of having the lowest value in one of the Euclidean space dimensions) of a intersection of $n$ convex sets is always the minimal point of the intersection of two convex sets in the family. Has this been proven before?

Adrian TC
  • 51
  • 1

1 Answers1

3

In $d$ dimensions, you need at least $d$ convex sets, not two. For example, in a suitably rotated polytope in $d$ dimensions, the minimal point will be a vertex, which is defined by the intersection of $d$ planes.

Geoffrey Irving
  • 3,969
  • 18
  • 41
  • Many thanks Geoffrey, my problem is in two dimensions. Is there a known result that states that in $d$ dimensions you need at least $d$ convex sets to define a minimal point? – Adrian TC Oct 31 '12 at 17:32