Experiments of Slammer worm propagation simulation in PDNS show that the optimized partitioning method can improve the performance of simulation by over 20% compared to the partitioning method based on graph partitioning tools.
As there are shortcomings of the partitioning method based on graph partitioning tools, a method for optimized partitioning for parallel simulation task of network worms is developed: By treating the model for estimating the running time of parallel simulation of network worms as optimization object function, the simulation task of worms is partitioned by the improved simulated annealing.
Shi and Malik converted image segmentation into graph partitioning problem and proposed the normalized cut criterion. Rather than focusing on local features and their consistencies in the image data, this approach aims at extracting the global impression of an image.
In this thesis we discuss issues around these problems, especially propositional satisfiability(SAT) and graph coloring, and different solving algorithms of them - the new born SAT algorithm Survey Propagation(SP), a graph partitioning scheme Part for graph coloring and different symmetry breaking SAT algorithms X-Zchaff for graph coloring.
(4) establishing a full graph, namely Scene Structure Graph (SSG), and testifying some important properties of the adjacency matrix of the SSG, designing a fast spectral graph partitioning algorithm, namely Iterative Spectral Graph Partitioning Algorithm;