Fast parallel algorithms for graph-theoretic problems: matching, coloring, and partitioning
New parallel algorithms are presented to solve graph-theoretic problems of three kinds: matching, coloring, and partitioning. Throughout, superfast algorithms, are sought, those running on a parallel random access machine in time polynomial in the log of the input size (polylog time) and using a polynomial number of processors. Problems solvable with such algorithms are said to be in NC. Those solvable by randomized algorithms obeying the same time and processor bounds are said to be in RNC or LVNC; those in RNC (or Monte Carlo RNC) are solvable by algorithms which, on instances of size n, return a correct answer with probability at least 1-2/sup -n/, and those in LVNC (or Las Vegas RNC), by algorithms that always return either a correct answer or failure, failure being returned at most half the time. Often the algorithms themselves will be said to be in NC, TNC, or LVNC.
- Research Organization:
- California Univ., Berkeley (USA)
- OSTI ID:
- 5062298
- Resource Relation:
- Other Information: Thesis (Ph. D.)
- Country of Publication:
- United States
- Language:
- English
Similar Records
Two processor scheduling
Applications of parallel scheduling to perfect graphs. Technical report