Let P(d,v) denote the number of convex polytopes (up to combinatorial equivalence) in $\,$d dimensions with v vertices. The octahedron and triangular prism are two familiar examples with d=3 and v=6 . There are five others bringing the total to P(3,6) = 7.
In four dimensions, any polytope with 6 vertices has just one more than the bare minimum needed to form a simplex. $\,$ Apparently there are P(4,6) = 4$\,$ such.
For completeness, there is only the hexagon if d=2 and only the simplex when d=5.$\,$ Therefore, the full count of all polytopes with v=6 and $2\le d$$\,$$\le 5$ $\,$is given by $\,$(1,7,4,1).
Questions: (1) Is there a reference which shows illustrations for the seven polytopes in the$\,$ d=3, v=6 case?
(2)$\,$The set of 3 x 3 doubly stochastic matrices forms a four dimensional polytope inside$\,$ $\Bbb R^9$$\,$having six vertices (corresponding to the six permutation matrices). Is there a nice way to embed this in$\,$$\,$$\Bbb R^4$$\,$ or to understand it in some more direct geometric way?
(3)$\,$Suppose we fix$\,$ v $\,$the number of vertices and consider the sequence P(d,v) for $\,$$2 \le d$$\,$$ \le v-1$. $\,$ Is this sequence known or conjectured to be unimodal?
Thanks very much
An answer to (2). Yes. Let $P=\|p_{ij}\|$ be a $3\times 3$ doubly stochastic matrix. The polytope belongs to the intersection of six hyperplanes:
$p_{11}+p_{12}+p_{13}=1$
$p_{21}+p_{22}+p_{23}=1$
$p_{31}+p_{32}+p_{33}=1$
$p_{11}+p_{21}+p_{31}=1$
$p_{12}+p_{22}+p_{32}=1$
$p_{13}+p_{23}+p_{33}=1$.
We can try to construct a basis in this intersection space and then describe the polytope by its coordinates with the given basis.
A more tricky but clear way is to consider a linear map $L$, $P\mapsto (p_{11},p_{13},p_{31},p_{33})$ . Since $P$ is a doubly stochastic matrix, the map $L$ is injective. Indeed, the elements of the matrix $P$ can be recovered from its image $L(P)$ as follows:
$p_{12}=1-p_{11}-p_{13}$,
$p_{32}=1-p_{31}-p_{33}$,
$p_{21}=1-p_{11}-p_{31}$,
$p_{23}=1-p_{13}-p_{33}$,
$p_{22}=1-p_{21}-p_{23}$.
The map $L$ maps the six permutation matrices to the points
$(1,0,0,1)$, $(1,0,0,0)$, $(0,0,0,1)$, $(0,0,1,0)$, $(0,1,0,0)$, and $(0,1,0,1)$.