Description: Novadays, the methods of computational intelligence are widely used to solve a problems related to information processing, including both traditional Data Mining and new directions like Dynamic Data Mining, Data Stream Mining, Big Data Mining, Web-Mining, Text Mining, etc. One of the main areas of computational intelligence are evolutionary algorithms, which are represented in the form of global models of the reproduction or development of biological organisms, which are intended, in general, in order to find a global optimum of multiextremal functions under conditions of uncertainty. The most popular evolutionary bioinspiral algorithms are the so-called Particle Swarm Optimization (PSO), among which the most promising both from the point of view of speed and ease of implementation are the optimization algorithms based on the Cat's Swarm (Cat Swarm Optimization - CSO). In this paper, we consider the optimization problem based on modified Cats Swarm approach by introducing random search elements into the seeking and tracing processes that permits to improve the position of results and provide global properties to tracing mode. The proposed optimization method, being the representative of evolutionary algorithms, is intended for use in hybrid systems of computation intelligence and, above all, in the tasks of learning artificial neural networks, neuro-fuzzy systems, as well as in the tasks of clusterization and classification.
Keywords: optimization, cats swarm, tracing mode, seeking mode