Multi-agent genetic algorithm for solving knowledge extraction problem

The paper deals with one of the tasks of data intellectual analysis - the task of knowledge extraction. The urgency of solving this problem is caused by the absence of formal models of the facilities and the need for a priori knowledge of the incoming data. To solve these problems neural network technology, methods of evolutionary modeling, genetic and other population algorithms are used. The article describes the advantages of using such methods, and the ways to enhance their effectiveness are proposed. A research of software environment, which implements proposed multi-agent approach was developed, and series of computational experiments were performed.