Семинар за рачунарство и примењену математику, 27. децембар 2016.

Наредни састанак Семинара биће одржан у уторак, 27. децембра 2016. у сали 301ф Математичког института САНУ са почетком у 14:15 часова.

Предавачи: Слободан Симић, Ранко Поповић, Математички институт САНУ, Далибор Ристић, Рачунарски факултет


Апстракт: The paper presents the main features of NeoGraph, the software platform for studying complex networks. It is to a large extent based on spectral graph theory, and is intended for both development of spectral graph theory and its application in solving theoretical and practical problems of computer science, especially in revealing phenomena which are intrinsic for large scale networks.

Spectral graph theory, as a mathematical discipline, makes use of linear algebra in studying graphs and various real networks. The basic tools are embodied in various matrices, whose spectral domain, i.e., eigenvalues and eigenvectors, provide powerful facilities. The most important benefit in using spectral techniques, from the practical point of view, comes from the fact that the numeric algorithms for computing eigenvalues and eigenvectors are polynomial, and also are considerably faster for sparse graphs which are one of basic features of complex networks. NeoGraph is primarily tailored for investigations involving complex networks in the context of modeling, simulation, synthesis, decomposition, searching, analysis, visualization, etc. However, the platform can also be used in education of students, in research (say, for making or disproving conjectures), or in solving real-life problems.

This is joint work with D. Živković, A. Ivanović, B. Milojković, A. Stančić