The structure of intelligent agent projecting and the possibilities of its execution

Most of today's information systems are not designed for independent decision making, behavior of such systems should be incorporated at the design stage. Entering the conditions that are not included by system developers, lead to a crash. The exponential growth of computing power of modern processors leads to an increase in range of tasks that can be automated, and an increase in their complexity. Partially testing helps to solve the problem, but it has a drawback, precisely the behavior of the system is usually tested in conditions which are already taken into account at the design stage. One approach to address this problem is the agent-oriented programming and the use of multi-agent systems. The aim of the work is to highlight contemporary approaches: a definition of agent concepts, applications of multi-agent systems, mathematical model agent, modern methods of design and implementation of intelligent agents, also covered approaches to the definition of the concept of an intelligent agent, the application of intelligent agents and multi-agent systems, and methods theoretically agent and describe its implementation. The basis of these methods of abstract description of the intelligent agent is the works of M.Wooldridge and N.Jennings. We propose several methods to describe such behaviors of intelligent agent as self-learning, goal setting and forecasting, and planning. Therefore we cover the concept of «intent» and widespread mental (beliefs-desires-intentions BDI) architecture of intelligent agent, as well as the proposed methods of their description in the framework of the developed formalism.