# Distance between two sets

Let $D(A ,B)= \inf_{\alpha \in A,\beta \in B} d(\alpha ,\beta )$ denote the distance between sets in the power set of a given set where $d(\alpha ,\beta )$ is a metric. Prove that $D(A,B )=0$ doesn't imply $A\bigcap B\neq \varnothing$.

My attempt: Let $D(A,B)= 0$. Then, $\inf_{\alpha \in A,\beta \in B}(d(\alpha ,\beta ))=0$. So there's $a\in A , b\in B$ such that $d(a,b)=0$, which means $a=b$. So, $A\bigcap B\neq \varnothing$. What's the problem in my proof?

-
I think you mean $D(A,B)=...$. It feels like you copied the problem wrong. – Thomas Andrews Dec 3 '12 at 17:01
Yes , Sorry for that. – naanwa Dec 3 '12 at 17:11
You still need to edit the other instances of $D(\alpha,\beta)$ to be $D(A,B)$. Specifically, "Prove that $D(\alpha,\beta)=0$..." – Thomas Andrews Dec 3 '12 at 17:12

There dont have to be such elements $a$ and $b$

for example take $$A = \left\{ (x,y) \,\vert\, x=0\right\}$$

and $$B = \left\{ (x,y) \,\vert\, x=1/y\right\}$$

-

The infimum of a set is not necessarily attained for any particular values of the set (so your claim "there is $a\in A$ and $b\in B$ such that $d(a,b)=0$" is an issue).

Hint: try to construct an explicit example of such sets, for instance in $\Bbb{R}$.

-

This statement “Then , $\inf_{\alpha\in A,\beta \in B}(d(\alpha,\beta))=0$. So there's $\alpha \in A, \beta \in B$.” is incorrect. It would be correct if we had $\min_{\alpha\in A,\beta \in B}(d(\alpha,\beta))=0$.

-

Take $A=[-1,0]$ and $B={1/n:n=1,2,...}$. Then $A$ and $B$ have no intersection but their distance is 0.

-
It's a good example, but when answering a three-year old Question, you would be helping your Readers to highlight any significant points raised by your Answer that are not already present in the ones posted previously. – hardmath Apr 18 at 3:42