The maximum number of faces of the Minkowski sum of two convex polytopes (Q309640)

From MaRDI portal





scientific article; zbMATH DE number 7053255
Language Label Description Also known as
default for all languages
No label defined
    English
    The maximum number of faces of the Minkowski sum of two convex polytopes
    scientific article; zbMATH DE number 7053255

      Statements

      The maximum number of faces of the Minkowski sum of two convex polytopes (English)
      0 references
      0 references
      0 references
      7 September 2016
      0 references
      10 May 2019
      0 references
      The paper studies the maximum number of \(k\)-faces the Minkowski sum of two \(d\)-dimensional polytopes can have for any \(k\in\{0,1,\ldots,d-1\}\). Firstly, the \textit{Cayley polytope} \(P\) of two \(d\)-polytopes \(P_1\) and \(P_2\) is the \((d+1)\)-dimensional polytope in \({\mathbb R}^{d+1}\) obtained by taking the convex hull of \(P_1\) embedded in the hyperplane \(x_1 = 0\) and \(P_2\) embedded in the hyperplane \(x_1 = 1\). Secondly, a polytope \(Q\) is said to be \textit{\((k,V_1)\)-bineighborly} if the vertex set \(V\) of \(Q\) has a partition \(V = V_1\cup V_2\) such that for every proper subsets \(S_1\subseteq V_1\) and \(S_2\subseteq V_2\) with \(|S_1|+|S_2|\leq k\), the union \(S_1\cup S_2\) is the vertex set of a \((|S_1|+|S_2|-1)\)-face of \(Q\). The main result of this paper is the following theorem: Let \(P_1\) and \(P_2\) be two \(d\)-polytopes with \(n_1\geq d+1\) and \(n_2\geq d+1\) vertices respectively. For any \(k\in\{1,\ldots,d\}\) the number \(f_{k-1}(P_1+P_2)\) of \((k-1)\)-faces of the Minkowski sum \(P_1+P_2\) is bounded from above in the following way \[ \begin{multlined} f_{k-1}(P_1+P_2) \leq f_k(C_{d+1}(n_1+n_2))\\ - \sum_{i=0}^{\lfloor (d+1)/2\rfloor}\binom{d+1-i}{k+1-i} \left(\binom{n_1-d-2+i}{i}+\binom{n_2-d-2+i}{i}\right), \end{multlined} \] where \(C_{d+1}(n_1+n_2)\) denotes the cyclic \((d+1)\)-polytope with \(n_1+n_2\) vertices. In addition, (i) equality holds for all \(k\in \{1,\ldots,l\}\) iff \(l\leq \lfloor (d-1)/2\rfloor\) and the Cayley polytope \(P\) is \((l+1,V_1)\)-bineighborly, (ii) for even \(d\geq 2\) equality holds for all \(k\in \{1,\ldots,d\}\) iff \(P\) is \(\lfloor d/2 \rfloor\)-neighborly and (iii) for odd \(d\geq 3\) equality holds for all \(k\in \{1,\ldots,d\}\) iff \(P\) is \((\lfloor (d+1)/2 \rfloor,V_1)\)-bineighborly. The proof is analogous to that of the Upper Bound Theorem by McMullen from 1970 [\textit{P. McMullen}, Mathematika 17, 179--184 (1970; Zbl 0217.46703)]. These bounds are all shown to be tight in the last section of the paper. The three figures in the paper are interesting, intricate and explanatory.
      0 references
      high-dimensional geometry
      0 references
      discrete geometry
      0 references
      combinatorial geometry
      0 references
      combinatorial complexity
      0 references
      Minkowski sum
      0 references
      convex polytopes
      0 references
      cs.CG
      0 references
      math.CO
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references