PARTY

PARTY: The problem of partitioning a graph into a number of pieces is one of the fundamental tasks in computer science and occurs in several applications like parallel programming or VLSI design.Finding optimal partitions according to different measures is in most cases NP-complete. A large number of efficient partitioning heuristics have been developed, but their performance in terms of computation time as well as solution quality is heavily influenced by choices of parameters and certain implementation details. Fortunately, the partitioning problem itself is clearly defined and its description leads to a small interface. The PARTY partitioning library serves a variety of different partitioning methods. Instead of implementing the methods by himself, the user may take advantage of the PARTY implementations. The performance of the code has been improved over several years and the default setting of the library enables a fast and easy start for the user. Two kinds of interfaces allow the use as stand-alone tool as well as the inclusion into application codes. Several research projects currently use the PARTY partitioning library successfully to solve their partitioning problem. For further information, in particular on how to obtain it, please contact Prof. Dr. Robert Preis.