The project deals with some currently 'hot topics' such as: data mining, link analysis, visualization, complex systems. Theoretical background and properties of several efficient procedures for analyzing large datasets and networks were developed: - addapted leaders' method for clustering of large sets of symbolic objects; - addapted hierarchical clustering method for symbolic objects, compatible with the leaders' method; - an O(m) algorithm for k-cores in networks; - an O(m log n) algorithm for generalized cores in networks; - a triadic census algorithm for large sparse networks; - an algorithm for 2-mode generalized blockmodelin; - algorithms for triangular connectivity and determining of triangular networks; - short-cycle connectivity and corresponding networks; - algorithms for determining vertex- and line-islands in networks. These algorithms were integrated in program Pajek. It is currently the only available general purpose program for analysis and visualization of large networks. It has a broad users community all over the world. Pajek is available at: http://vlado.fmf.uni-lj.si/pub/networks/pajek/ We developed also several efficient algorithms for different combinatorial optimization problems.