Modeling of resources management system of heterogeneous distribution system of information processing on the basis of multiagent approach

The article is devoted to the structural modeling of resource management of heterogeneous distribution systems of information processing based on multi-agent approach of the design of distributed information systems. In the introduction of the article the urgency of the problem for the decision-making process is proved. We describe the main difficulties encountered when we make informed and timely decisions at forecasting and disaster management on the scale of the Russian Federation. Further objective of the study is formed, and the tasks required to achieve it are described. Starting with a description of the distinguishing features of the unified state system of prevention and liquidation of emergency situations of the Russian Emergencies Ministry, the authors come to the conclusion that the system fits the generally accepted concept of heterogeneous distributed information processing system. Below, the construction of multi-agent approach to resource management of heterogeneous distributed information processing system as the least costly financially, and meets the necessary criteria is justified. In the body of the article a complete description of levels built models such as the – level of intelligent agents, the level of computer equipment agents, route management level is given. The structure of the system is described. The article details the typical functionality of the following: a software component management agent compute node load monitoring agent and monitoring agent communication channels, data transfer agent, load balancing agent, agent utilization of communication channels forecasting reactive agent of the primary data sources. At the end of the work the possibilities of further development of the model, as well as the conclusion about the necessity of its algorithmization for practical testing and implementation in practice are considered.

Keywords: modeling, heterogeneous data processing systems, multi-agent systems.

Full text of the article