Macrosystems Models of Dynamic Stochastic Networks

Abstract

The state-of-the-art of the worldwide web (Internet) and its short-range development (GRID), on the one hand, offer to the user wide possibilities of accessing not only the international information resources, but the computational resources as well, and, on the other hand, give rise to a number of problems of system nature. GRID was considered as a dynamic stochastic network whose state is characterized by the spatial distributions of the information-and-computation resources and information flows. The time profiles of these components of state feature essentially different times of relaxation. This distinction of the network enables one to adapt and develop the concepts of the macrosystem approach for studying the space-time evolution of the GRID state. Proposed were models of the states of the local-stationary network that rely on the generalized principle of entropy maximization. They make an integral part of the evolutionary model of information-and-computation resources which is classified with the positive dynamic systems with an entropy operator.