Using out-of-core techniques to produce exact solutions to the maximum clique problem on extremely large graphs

Practical methods are presented for computing exact solutions to the maximum clique problem on graphs that are too large to fit within core memory. These methods use a combination of in-core and out-of-core techniques, recursively dissecting large graphs into manageable components. A global solution to the maximum clique problem is derived from local solutions generated for each of the individual components. Parallelizing the search within these components is instrumental in improving the running times of the algorithms.