A New Approach to Exact Crossing Minimization

The crossing number problem is to find the smallest number of edge crossings necessary when drawing a graph into the plane. Eventhough the problem is NP-hard, we are interested in practically efficient algorithms to solve the problem to provable optimality. In this paper, we present a novel integer linear programming (ILP) formulation for the crossing number problem. The former formulation [4] had to transform the crossing number polytope into a higher-dimensional polytope. The key idea of our approach is to directly consider the natural crossing number polytope and cut it with multiple linear-ordering polytopes. This leads to a more compact formulation, both in terms of variables and constraints. We describe a Branch-and-Cut algorithm, together with a combinatorial column generation scheme, in order to solve the crossing number problem to provable optimality. Our experiments show that the new approach is more effective than the old one, even when considering a heavily improved version...

• Type: Conference
• Year: 2008
• Where: ESA
• Authors: Markus Chimani, Petra Mutzel, Immanuel M. Bomze
• Keywords: Algorithms, ESA 2008, Number Polytope, Number Problem, Provable Optimality