The problem of exact symmetry detection in general graphs has received much attention recently. In spite of its NP-hardness, two different algorithms have been presented that in general can solve this problem quickly in practice [5,2]. However, as most graphs do not admit any exact symmetry at all, the much harder problem of fuzzy symmetry detection arises: a minimal number of certain modifications of the graph should be allowed in order to make it symmetric. We present a general approach to this problem: we allow arbitrary edge deletions and edge creations; every single modification can be given an individual weight. We apply integer programming techniques to solve this problem exactly or heuristically and give runtime results for a first implementation.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
- An Integer Programming Approach to Fuzzy Symmetry Detection
- Springer Berlin Heidelberg
ec4u, Neuer Inhalt/© ITandMEDIA