Define the set:

$$X:=\{K \subset \mathbb{C}: K \text{ is bounded and closed } \}.$$

Define a function $d : X \times X \rightarrow \mathbb{R}$ via:

$$d(K_1,K_2) = \inf \{ \delta > 0 : K_1 \subset N_{\delta}(K_2) \text{ and } K_2 \subset N_{\delta}(K_1)\},$$


$$N_δ(K) = \{ x \in \mathbb{C} : \exists y \in K \text{ with } |x−y|< \delta \}.$$

Note that we could also write $N_{\delta}(K) = \bigcup \limits_{y \in K} N_{\delta}(y)$, where $N_{\delta}(y) = \{x \in \mathbb{C} : |x-y|<\delta\}$.

I'm having trouble with understanding $d$ and $N_δ(K)$. Can someone explain me why $d(K_1,K_2)$ is a metric?

  • $\begingroup$ I have edited your post (it is currently being peer reviewed); please check that the edits correspond with your original intent. $\endgroup$ – Sambo Oct 13 '18 at 18:03

Intuitively $N_{\delta}$ takes a set $K$ and 'fattens' it by radius of $\delta$. An alternative way of writing it would be $$N_\delta(K) = K + B_{\delta}(0) $$ $d$ then takes two sets $K_1$ and $K_2$, and measures how much we need to 'fatten' $K_1$ until it contains all of $K_2$ and vice versa. Another way of writing it is $$ d(K_1, K_2) = \inf \{\delta > 0 : (\forall x \in K_1)(\exists y \in K_2)\ \lvert x - y \rvert \leq \delta \ \text{and}\ (\forall y \in K_2)(\exists x \in K_1)\ \lvert x - y \rvert < \delta\} $$ $d$ is well defined since the sets $K_1$ and $K_2$ are bounded, thus we aren't taking the infinum over an empty set.

By $d$'s very definition, it is symmetric and positive. What remains is to show $d(K_1, K_2) = 0 \iff K_1 = K_2$ (this requires the compactness of $K_1$ and $K_2$) and the triangle inequality.

  • 1
    $\begingroup$ It's worth nothing that the sets $K_1$ and $K_2$ should also be non-empty for $d$ to be well-defined! $\endgroup$ – Sambo Oct 13 '18 at 18:15
  • $\begingroup$ @bitesizebo if $d(A,B) = 0$, then every point of A is within zero distance of B,and hence must belong to B since B is compact, so $A \subset B$ and similarly $B \subset A$. So $d(A, B) = 0$ implies that $A = B$. Is this it for definiteness? And how do I go with the triangle inequality? $\endgroup$ – dxdydz Oct 14 '18 at 14:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.