Lower bound for arrangement of hyperplanes

76 Views Asked by At

There are many questions conserning an upper bound on the number of regions in an arrangement of hyperplanes in general position, but I'm interested in a lower bound for hyperplanes with some extra conditions.

Let $A_1,...,A_k$ be sets of each $s$ parallel hyperplanes in $\mathbb{R}^n$ and let $d(n,k,s)$ denote the minimum number of regions in such an arangement. We assume that hyperplanes in different $A_i$ are not parallel. Do we know a lower bound to this?

I think $d(n,k,s)=d(2,k,s)$ holds. Further it is clear that $d(n,1,s)=s+1$.

Any idea or source for some more information would helps. Best regards and many thanks in advance.

Here are my own thoughts in dimension two: The parallels from $A_1$ divide the plane in $s+1$ regions. Adding a line from $A_2$ splits this regions, so after adding all the lines form $A_2$ we get $(s+1)+ s \cdot (s+1)$ regions. This is best we can get also for the over $A_i$, so in conclusion we get $(s+1)+ (k-1) \cdot s \cdot (s+1)$ regions. Question: Is this optimal? What if we in dimension 3 we assume that not all planes contain one parallel direction?