Design of grid systems to solve complex industrial engineering problems

The problem of solving complex and demanding tasks of production engineering is discussed. The complexity of the methods used in the solution of such tasks requires significant computational resources. It is proposed to use the widespread personal computers combined into a single computer network on the basis of the Grid technology. However, the informal aspect is the choice of an effective structure of such a system. A complex of mathematical models for assessing performance and reliability of Grid systems and the corresponding problem of selecting its effective structure is proposed. On the basis of the proposed models the decision support system for the Grid systems formation is developed. To solve optimization problems arising in the process of decision-making, the multiobjective genetic algorithm is offered. When studying the proposed system efficiency, a practical problem of designing effective grid system configuration was solved, which is focused on the complex task of network planning by means of artificial neural networks.