Nasazení algoritmů pro dělení rozsáhlých grafů na cluster pomocí MPI.

Abstract

This work is focused on method called random walk. It contains explanation of basic definitions in field of graphs concerning graph partitioning. Further attention is then given to some methods of clustering of big graphs. Random walk method will be also implemented in two ways - serial and paralel. Paralel implementation will be based on MPI library. Implemented algorithms were tested from different points of view and quality of clustering was also measured using several measuring methods.