Number of intersections in a family of compact sets that cover the Euclidean $d-$plane.

40 Views Asked by At

I am in particular interested in families of compact sets given by $\{K+z_j:j\in \mathcal{J} \}$ such that $\mathbb{R}^d = \bigcup_{j\in \mathcal{J}}K+z_j$ where $K\subset \mathbb{R}^d$ is compact and $\{z_j:j\in \mathcal{J} \}\subset\mathbb{R}^d$ is just a sequence of points in $\mathbb{R}^d.$ How do I show that $$ \sup_{x\in \mathbb{R}^d} \# \{j\in\mathcal{J}: x\in K+z_j \} = \sup_{i\in \mathcal{J}} \# \{i\in\mathcal{J}: (K+z_i) \cap (k+z_j) \neq \emptyset \}? $$

For $x\in \mathbb{R}^d,$ $i\in \mathcal{J}$ we can define $B_x := \{j\in\mathcal{J}: x\in K+z_j \}$ and $Q_i:= \{i\in\mathcal{J}: (K+z_i) \cap (k+z_j) \neq \emptyset \}$. I was able to show that $\sup_{x\in \mathbb{R}^d} \#B_x \leq \sup_{i\in \mathcal{J}}\# Q_i$ by noting that for any $x\in \mathbb{R}^d$ we can always find an $i\in \mathcal{J}$ such that $x\in K+z_i,$ so $\#B_x \leq \#Q_i$. The reverse inequality remains elusive to me, and I suspect it has something to do with the compactness of $K$. Anybody knows if this is actually doable without extra hypotheses? If it is, can you give me some hints?

For some context, I am studying Amalgam Spaces (Introduction to Weighted Wiener Amalgams by C. Heil), and this equality came up without proof in the definition of BUPUs.

1

There are 1 best solutions below

0
On

It is not true. In fact we always have $\sup_{x\in \mathbb{R}^d} \#B_x \leq \sup_{i\in \mathcal{J}}\# Q_i$. This is true because $\sup_{x \in K + z_i} \#B_x \le \#Q_i$ so that

$$\sup_{x\in \mathbb{R}^d} \#B_x = \sup_{i \in \mathcal{J}} \sup_{x \in K + z_i} \#B_x \le \sup_{i \in \mathcal{J}} \#Q_i .$$

Now consider the following example for $d=2$. Let $K = [0,1] \times [0,1]$ and let $z_j$ be all points having the form $(m,n)$ with $m, n \in \mathbb{Z}$ and $m$ even and $(m,n+\frac{1}{2})$ with $m, n \in \mathbb{Z}$ and $m$ odd. Then it easy to see that $\sup_{x\in \mathbb{R}^d} \#B_x = 3$ and $\sup_{i \in \mathcal{J}} \#Q_i = 7$.