PARALLEL ALGORITHMS FOR COMPLEX NETWORKS SIMULATIONS
ITMO University, Saint Petersburg, 197101, Russian Federation ; Associate professor
I. . Kolyhmatov
ITMO University; student
A. V. Boukhanovsky
ITMO University, Saint Petersburg, 197101, Russian Federation; Director
Abstract. The set of parallel algorithms which allow reproducing a complex network as a stochastic graph with probabilistic properties is proposed in the paper. Each of algorithms is based on natural properties of graph structure parallelization and corresponds to features of the certain parallel computer architecture, including systems with the shared and distributed memory, and cluster systems as well. Results of experiments of parallel efficiency on hierarchical parallel computer systems and cluster systems (T-Forge mini) are presented.
Keywords: complex networks, generation the graph, parallel algorithms, hybrid architecture, models of productivity.