The set of all nonempty and convex subsets of $\mathbb{R}$ is a vector space...

286 Views Asked by At

...where the addition and scalar multiplication is defined as $A+B=\{a+b|a\in A \ and \ b\in B\}$ and $\alpha . A = \{\alpha a | a \in A\}$, where $\alpha \in \mathbb{R}$.

I think this claim is false, since $A=[0,1]$ and $-A=[-1,0]$ are convex subsets of $\mathbb{R}$ but $A+(-A)\neq {0}$. Is my reasoning correct?

Is it possible to create a vector space with intervals or non-discrete sets?

Thank you in advance for any tip and advice.

1

There are 1 best solutions below

3
On

Your reasoning is correct, though you might first want to verify that the zero element with respect to this addition is the nonempty convex subset $\{0\}\subset\Bbb{R}$.

Then it is indeed the case that $$A+(-A)=\{a+a':\ a\in[0,1],\ a'\in[-1,0]\}\neq\{0\},$$ because $1+0=1$ and $0+(-1)=-1$ are elements of $A+(-A)$, but not of $\{0\}$.

Edit: As noted in the comments, you might also want to verify that no other nonempty convex subset of $\Bbb{R}$ is the inverse of $A$, i.e. no other nonempty $B\subset\Bbb{R}$ satisfies $$A+B=B+A=\{0\}.$$ Of course $A\subset A+B$ so all convex nonempty subsets of $\Bbb{R}$ fail to have an inverse, except $\{0\}$ itself.

Alternatively, if the set is a vector space then multiplication distributes over addition and so $$A+(-A)=1\cdot A+(-1\cdot A)=(1+(-1))\cdot A=0\cdot A=\{0\},$$ which we have just shown to be false.