Is $\mathbb{R}^2$ a convex set without extreme points?

673 Views Asked by At

My reasoning is, the vector space being a 'flat plane', a straight line joining any two points on it will always be contained in $\mathbb{R}^2$. And it has no points in it which cannot be on the straight line connecting two other points in the set (since the plane is infinite)

I am new to convex analysis and optimization techniques, so my line of thinking is very informal. Please let me know whether or not my approach is correct.

2

There are 2 best solutions below

9
On BEST ANSWER

For any point $E \in \mathbb{R}^2$, because $\mathbb{R}^2$ is closed under addition and scalar multiplication, there exist two points (vectors) $x,y \in \mathbb{R}^2$ and an $\alpha \in (0,1)$ such that $E = \alpha x + (1-\alpha)y \in \mathbb{R}^2$. Thus, the space does not contain any extreme points.

8
On

You're right that it is convex. Taking any $\vec x,\vec y\in\Bbb R^2,$ we have for all $\alpha\in(0,1)$ that $\alpha\vec x,(1-\alpha)\vec y\in\Bbb R^2$ by closure under scalar multiplication, and so $\alpha\vec x+(1-\alpha)\vec y\in\Bbb R^2$ by closure under addition.

As for extreme points, note that if $\vec x\ne\vec 0,$ then $\vec x$ lies on the open line segment between $\frac12\vec x$ and $\frac32\vec x,$ so cannot be extreme. I will leave it to you to show (by example) that $\vec 0$ cannot be an extreme point, either.