Extended Formulations with small Diameter

Graphs of polytopes form an important part of polytope theory, since the graph encodes a lot about the combinatorial structure of the polytope. In particular, bounding the diameter of the graph of a polytope is a well-known and long-standing open problem. A great theoretical and practical interest in this question originates from the desire to efficiently solve Linear Programming, since the existence of a polynomial (in the number of polytope facets) bound on the diameter is a necessary condition for the yet undecided existence of a polynomial time variant of the Simplex-Algorithm.

However, in order to efficiently optimize linear functions over a polytope P, it suffices to efficiently optimize linear functions over an extension Q of P. The latter observation leads to the following question we call Extended Hirsch-Conjecture: Is there a polynomial function q(ยท) such that every simple polytope P with n facets admits a simple extension Q with at most q(n) facets and diameter at most q(n)?.

This project aims at investigating the above formulated Extended Hirsch-Conjecture, with the goal of collecting results that provide a better understanding of whether allowing for extensions indeed helps in proving small bounds on diameters or whether that problem appears not to be relaxed significantly by this approach.

currently no upcoming news
...more
currently no upcoming news
...more